MCP Tools

For connection setup, tool discovery, and transport readiness, an agent uses this public tool to check the state or evidence for "Establish MCP connection". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For connection setup, tool discovery, and transport readiness, an agent uses this public tool to check the state or evidence for "Tool-Discovery". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For connection setup, tool discovery, and transport readiness, an agent uses this public tool to check the state or evidence for "App-Info read". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For connection setup, tool discovery, and transport readiness, an agent uses this public tool to check the state or evidence for "Instructions fetch". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For connection setup, tool discovery, and transport readiness, an agent uses this public tool to handle the sensitive path "Permission Prompt". It keeps purpose, scope, user control, and sensitive data handling explicit.

For connection setup, tool discovery, and transport readiness, an agent uses this public tool to carry out the bounded step "App-Info Build/Runtime Diagnostics". It separates target confirmation, intended state change, and verification after execution.

For connection setup, tool discovery, and transport readiness, an agent uses this public tool to check the state or evidence for "Instructions Framework Hints". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For connection setup, tool discovery, and transport readiness, an agent uses this public tool to start or monitor the bounded flow "Instructions Task Awareness". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For connection setup, tool discovery, and transport readiness, an agent uses this public tool to check the state or evidence for "Instructions Service Discovery". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For connection setup, tool discovery, and transport readiness, an agent uses this public tool to check the state or evidence for "Instructions Learn-Mode Progress". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For connection setup, tool discovery, and transport readiness, an agent uses this public tool to carry out the bounded step "Instructions Operator Notes". It separates target confirmation, intended state change, and verification after execution.

For connection setup, tool discovery, and transport readiness, an agent uses this public tool to check the state or evidence for "Bundle Taxonomie & Gating". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For connection setup, tool discovery, and transport readiness, an agent uses this public tool to check the state or evidence for "Capability Bootstrap Guidance". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For connection setup, tool discovery, and transport readiness, an agent uses this public tool to check the state or evidence for "Streamable-HTTP Handshake". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

Transport Auth/Origin/Loopback is a public reference for connection setup, tool discovery, and transport readiness. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For connection setup, tool discovery, and transport readiness, an agent uses this public tool to check the state or evidence for "Konkrete Unavailable-Tool-Gates". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For connection setup, tool discovery, and transport readiness, an agent uses this public tool to check the state or evidence for "MCP-Transportlog read". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For tabs, active targets, and claim state, an agent uses this public tool to check the state or evidence for "Open tab". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For tabs, active targets, and claim state, an agent uses this public tool to carry out the bounded step "Tab claimen". It separates target confirmation, intended state change, and verification after execution.

For tabs, active targets, and claim state, an agent uses this public tool to carry out the bounded step "Tab wechseln". It separates target confirmation, intended state change, and verification after execution.

For tabs, active targets, and claim state, an agent uses this public tool to check the state or evidence for "Tab-Liste". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For tabs, active targets, and claim state, an agent uses this public tool to check the state or evidence for "Tab-Snapshot". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For tabs, active targets, and claim state, an agent uses this public tool to carry out the bounded step "Tab releasen". It separates target confirmation, intended state change, and verification after execution.

For tabs, active targets, and claim state, an agent uses this public tool to carry out the bounded step "Close tab". It separates target confirmation, intended state change, and verification after execution.

Claim-Conflict is a public reference for tabs, active targets, and claim state. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For tabs, active targets, and claim state, an agent uses this public tool to carry out the bounded step "Reclaim". It separates target confirmation, intended state change, and verification after execution.

For tabs, active targets, and claim state, an agent uses this public tool to check the state or evidence for "Tab Activity Resume". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For tabs, active targets, and claim state, an agent uses this public tool to carry out the bounded step "Active-Tab Activation Edge". It separates target confirmation, intended state change, and verification after execution.

For tabs, active targets, and claim state, an agent uses this public tool to carry out the bounded step "Tab-New Auto-Claim". It separates target confirmation, intended state change, and verification after execution.

For tabs, active targets, and claim state, an agent uses this public tool to carry out the bounded step "Release Finalize Stats". It separates target confirmation, intended state change, and verification after execution.

Release Token Override is a public reference for tabs, active targets, and claim state. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For navigation, history, and controlled page changes, an agent uses this public tool to carry out the bounded step "Einfache Navigation". It separates target confirmation, intended state change, and verification after execution.

For navigation, history, and controlled page changes, an agent uses this public tool to carry out the bounded step "SPA-Route". It separates target confirmation, intended state change, and verification after execution.

For navigation, history, and controlled page changes, an agent uses this public tool to carry out the bounded step "Route with selector". It separates target confirmation, intended state change, and verification after execution.

For navigation, history, and controlled page changes, an agent uses this public tool to carry out the bounded step "Back/Forward". It separates target confirmation, intended state change, and verification after execution.

For navigation, history, and controlled page changes, an agent uses this public tool to carry out the bounded step "Reload". It separates target confirmation, intended state change, and verification after execution.

For navigation, history, and controlled page changes, an agent uses this public tool to carry out the bounded step "SPA-Session-Block". It separates target confirmation, intended state change, and verification after execution.

For navigation, history, and controlled page changes, an agent uses this public tool to carry out the bounded step "Force-Navigate". It separates target confirmation, intended state change, and verification after execution.

Auth-Persistence-Gate is a public reference for navigation, history, and controlled page changes. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For navigation, history, and controlled page changes, an agent uses this public tool to carry out the bounded step "Navigate Commit-Signal". It separates target confirmation, intended state change, and verification after execution.

For navigation, history, and controlled page changes, an agent uses this public tool to carry out the bounded step "SPA Back/Forward Same-Document". It separates target confirmation, intended state change, and verification after execution.

For navigation, history, and controlled page changes, an agent uses this public tool to check the state or evidence for "History Snapshot". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For navigation, history, and controlled page changes, an agent uses this public tool to check the state or evidence for "History Jump per EntryId". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For navigation, history, and controlled page changes, an agent uses this public tool to check the state or evidence for "History Jump per Offset". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

History No-op & Validation is a public reference for navigation, history, and controlled page changes. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

History Multi-Step Gate is a public reference for navigation, history, and controlled page changes. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

History Auth-Persistence is a public reference for navigation, history, and controlled page changes. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to check the state or evidence for "Perceive". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to check the state or evidence for "Page-Info". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to check the state or evidence for "DOM read". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to check the state or evidence for "Text read". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to check the state or evidence for "Screenshot". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to check the state or evidence for "Resource read". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to check the state or evidence for "Text search". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to check the state or evidence for "Console read". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to check the state or evidence for "Element-Rect". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to check the state or evidence for "Structured text". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to check the state or evidence for "Perceive Snapshot Query". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

Identity Overlay Warning is a public reference for page perception, diagnostics, and verifiable read signals. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to check the state or evidence for "Form-Analysis Autofill/Framework". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to check the state or evidence for "Resource Structured Bodies". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to check the state or evidence for "Essential Safety Fields". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to check the state or evidence for "Page-Read-Debug Bundle". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to check the state or evidence for "Resource Grep". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to check the state or evidence for "Network Read". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to check the state or evidence for "Messages Read". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Click". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Type + Verify". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Select Option". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Choose Option". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Scroll Smart". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Dismiss Blockers". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to handle the sensitive path "File Upload". It keeps purpose, scope, user control, and sensitive data handling explicit.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Keyboard-Shortcut". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Input-Text (raw)". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Input-Key (raw)". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Input-Click (raw)". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Input-Move (raw)". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to check the state or evidence for "Wait for Modal". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Wait for Selector". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Wait for Eval". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Autofill Popup Warning". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Type Truncation Warning". It separates target confirmation, intended state change, and verification after execution.

Native Dialog Warning is a public reference for targeted browser interaction with a prechecked target. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

File Input Risk Auto-Learn is a public reference for targeted browser interaction with a prechecked target. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Pre-Loop Stability Check". It separates target confirmation, intended state change, and verification after execution.

Aggressive Overlay Dismiss Telemetry is a public reference for targeted browser interaction with a prechecked target. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For targeted browser interaction with a prechecked target, an agent uses this public tool to handle the sensitive path "CTA Ref Selector Identity". It keeps purpose, scope, user control, and sensitive data handling explicit.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "CMP Apply". It separates target confirmation, intended state change, and verification after execution.

For risk-bounded actions with visible user control, an agent uses this public tool to check the state or evidence for "Guarded Send". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For risk-bounded actions with visible user control, an agent uses this public tool to handle the sensitive path "Guarded Login". It keeps purpose, scope, user control, and sensitive data handling explicit.

For risk-bounded actions with visible user control, an agent uses this public tool to carry out the bounded step "Guarded Submit". It separates target confirmation, intended state change, and verification after execution.

For risk-bounded actions with visible user control, an agent uses this public tool to carry out the bounded step "Guarded Switch Model". It separates target confirmation, intended state change, and verification after execution.

For risk-bounded actions with visible user control, an agent uses this public tool to carry out the bounded step "Guarded Switch Sandbox". It separates target confirmation, intended state change, and verification after execution.

For risk-bounded actions with visible user control, an agent uses this public tool to carry out the bounded step "GuardedSend Learning Suppression & Claim". It separates target confirmation, intended state change, and verification after execution.

For risk-bounded actions with visible user control, an agent uses this public tool to check the state or evidence for "Form Submission Bundle". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Proxy create". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Proxy-Liste". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Proxy-Status". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Proxy wechseln". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Proxy testen". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Proxy-Passwort set". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Browser-Web-Proxy-Log". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Proxy updaten". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Proxy disconnecten". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Proxy reconnecten". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Proxy delete". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Proxy Credentials Awareness Gate". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Proxy Bundle Discovery". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Vault save". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Vault read". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Secret tippen". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Vault delete". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Vault-Auth Bundle". It keeps purpose, scope, user control, and sensitive data handling explicit.

For bounded website review runs and discoverable result data, an agent uses this public tool to carry out the bounded step "Crawl start". It separates target confirmation, intended state change, and verification after execution.

For bounded website review runs and discoverable result data, an agent uses this public tool to start or monitor the bounded flow "Crawl-Status". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For bounded website review runs and discoverable result data, an agent uses this public tool to start or monitor the bounded flow "Crawl-Results". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For bounded website review runs and discoverable result data, an agent uses this public tool to carry out the bounded step "Crawl stop". It separates target confirmation, intended state change, and verification after execution.

For bounded website review runs and discoverable result data, an agent uses this public tool to check the state or evidence for "Site-URL-Index". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For bounded website review runs and discoverable result data, an agent uses this public tool to start or monitor the bounded flow "Crawl-History". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For bounded website review runs and discoverable result data, an agent uses this public tool to start or monitor the bounded flow "Crawl-Diff". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For bounded website review runs and discoverable result data, an agent uses this public tool to start or monitor the bounded flow "Crawl-Links". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For bounded website review runs and discoverable result data, an agent uses this public tool to carry out the bounded step "Crawl updaten". It separates target confirmation, intended state change, and verification after execution.

For bounded website review runs and discoverable result data, an agent uses this public tool to start or monitor the bounded flow "Crawl verifizieren". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For bounded website review runs and discoverable result data, an agent uses this public tool to check the state or evidence for "Site-URLs-Report". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For bounded website review runs and discoverable result data, an agent uses this public tool to carry out the bounded step "Discovery Reset". It separates target confirmation, intended state change, and verification after execution.

For bounded website review runs and discoverable result data, an agent uses this public tool to check the state or evidence for "Site-Discovery Probe". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For bounded website review runs and discoverable result data, an agent uses this public tool to handle the sensitive path "Site-Discovery Cache Read". It keeps purpose, scope, user control, and sensitive data handling explicit.

For bounded website review runs and discoverable result data, an agent uses this public tool to check the state or evidence for "MCP Inspect". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For bounded website review runs and discoverable result data, an agent uses this public tool to carry out the bounded step "MCP Connect Request". It separates target confirmation, intended state change, and verification after execution.

Drift Detection is a public reference for bounded website review runs and discoverable result data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For bounded website review runs and discoverable result data, an agent uses this public tool to check the state or evidence for "Instructions Discovery Digest". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For bounded website review runs and discoverable result data, an agent uses this public tool to check the state or evidence for "PKS Discovery Persistenz". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

Crawler Early-Probe & llms Seeds is a public reference for bounded website review runs and discoverable result data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For bounded website review runs and discoverable result data, an agent uses this public tool to start or monitor the bounded flow "Crawl Links Agent Signals". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

SiteArchitecture Automation Surface is a public reference for bounded website review runs and discoverable result data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Crawl Start SiteUrlIndex is a public reference for bounded website review runs and discoverable result data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Target Session Preservation Gate is a public reference for bounded website review runs and discoverable result data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Target Auth-Wall Gate is a public reference for bounded website review runs and discoverable result data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Finalizing Linger is a public reference for bounded website review runs and discoverable result data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Pause Draining is a public reference for bounded website review runs and discoverable result data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For bounded website review runs and discoverable result data, an agent uses this public tool to start or monitor the bounded flow "Crawl Verify URL Strictness". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

Crawl Verify Persisted Resume is a public reference for bounded website review runs and discoverable result data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Crawler Disabled Gate is a public reference for bounded website review runs and discoverable result data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For bounded website review runs and discoverable result data, an agent uses this public tool to carry out the bounded step "Discovery Reset Active Jobs". It separates target confirmation, intended state change, and verification after execution.

For bounded website review runs and discoverable result data, an agent uses this public tool to start or monitor the bounded flow "Crawler Bundle Discovery". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Phenomenon create". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Phenomenon read". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Phenomenon matchen". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Phenomenon patchen". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "PKS-Liste". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Phenomenon Apply". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Phenomenon deprecaten". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Domain-Hint". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "PKS Summary Output". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "PKS Route Scope". It separates target confirmation, intended state change, and verification after execution.

PKS Verified Health Fields is a public reference for knowledge signals, learning candidates, and explainable orientation. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

PKS Kontext-Advice is a public reference for knowledge signals, learning candidates, and explainable orientation. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Learn Suggest". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Learn Generate". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "Learn Promote". It separates target confirmation, intended state change, and verification after execution.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "Learn Feedback". It separates target confirmation, intended state change, and verification after execution.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "Revalidate". It separates target confirmation, intended state change, and verification after execution.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Telemetry Report". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

Semantic Learning Consent-CMP is a public reference for knowledge signals, learning candidates, and explainable orientation. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Semantic Learning Login-Wall is a public reference for knowledge signals, learning candidates, and explainable orientation. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Semantic Learning Risky-Repeat is a public reference for knowledge signals, learning candidates, and explainable orientation. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "Learning Opportunity Resolve". It separates target confirmation, intended state change, and verification after execution.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "Semantic Upsert Auto-Resolve". It separates target confirmation, intended state change, and verification after execution.

Semantic Advice Suppression is a public reference for knowledge signals, learning candidates, and explainable orientation. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Semantic Batch Summary is a public reference for knowledge signals, learning candidates, and explainable orientation. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Semantic Rate Limit pro Kind/Origin is a public reference for knowledge signals, learning candidates, and explainable orientation. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Semantic Priority Login vor CMP vor Repeat is a public reference for knowledge signals, learning candidates, and explainable orientation. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "FR-1 Selector Learning Debt". It separates target confirmation, intended state change, and verification after execution.

Learning Debt Auto-Upsert is a public reference for knowledge signals, learning candidates, and explainable orientation. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Learning Debt Regret". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "Source Opportunity Auto-Resolve". It separates target confirmation, intended state change, and verification after execution.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Semantic Metrics". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

Semantic Evidence Privacy is a public reference for knowledge signals, learning candidates, and explainable orientation. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Login-Wall MFA Signal is a public reference for knowledge signals, learning candidates, and explainable orientation. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

FR-2 supersedes FR-1 Debt is a public reference for knowledge signals, learning candidates, and explainable orientation. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

LearningRepairQueue Summary is a public reference for knowledge signals, learning candidates, and explainable orientation. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Durable AutoRepair". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "Learn Feedback Gate Breakdown". It separates target confirmation, intended state change, and verification after execution.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Learn-Onboarding Recall". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "Learn-Onboarding Confirm Gate". It separates target confirmation, intended state change, and verification after execution.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "PKS-Learning Bundle". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Learn-Onboarding Bundle-Membership". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "OK Observe". It separates target confirmation, intended state change, and verification after execution.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Perceive with PKS". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Form-Analysis". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "CTA-Detection". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

CTA Delta Fingerprint is a public reference for knowledge signals, learning candidates, and explainable orientation. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "OK Hints in Perceive". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to start or monitor the bounded flow "Task search". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to handle the sensitive path "Profil create". It keeps purpose, scope, user control, and sensitive data handling explicit.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "Instanz start". It separates target confirmation, intended state change, and verification after execution.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to start or monitor the bounded flow "Fortschritt melden". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to start or monitor the bounded flow "Task abclose". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to start or monitor the bounded flow "Task verifizieren". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "Guidance loggen". It separates target confirmation, intended state change, and verification after execution.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to start or monitor the bounded flow "Task matchen". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to handle the sensitive path "Profil read". It keeps purpose, scope, user control, and sensitive data handling explicit.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to handle the sensitive path "Alle Profile". It keeps purpose, scope, user control, and sensitive data handling explicit.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to start or monitor the bounded flow "Instanz Details". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to start or monitor the bounded flow "Promotion-Candidates". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "Guidance promoten". It separates target confirmation, intended state change, and verification after execution.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to start or monitor the bounded flow "Task Match Ranking Contract". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "Task Instance Effective Context". It separates target confirmation, intended state change, and verification after execution.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to start or monitor the bounded flow "Task Instance Resume Contract". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to start or monitor the bounded flow "Guidance Log Match Telemetry". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

Bootstrap-Warning is a public reference for agent warnings, blocks, and current approval signals. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Emergency Stop is a public reference for agent warnings, blocks, and current approval signals. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Disk-Space-Low is a public reference for agent warnings, blocks, and current approval signals. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Reflection Gate is a public reference for agent warnings, blocks, and current approval signals. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Semantic Learning Warn Gate is a public reference for agent warnings, blocks, and current approval signals. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For agent warnings, blocks, and current approval signals, an agent uses this public tool to carry out the bounded step "Semantic Learning Strict Gate". It separates target confirmation, intended state change, and verification after execution.

Semantic Gate Toggle Off is a public reference for agent warnings, blocks, and current approval signals. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

LearningGate Mode Matrix is a public reference for agent warnings, blocks, and current approval signals. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Perceive-First Gate is a public reference for agent warnings, blocks, and current approval signals. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For agent warnings, blocks, and current approval signals, an agent uses this public tool to check the state or evidence for "Tab-Awareness Warning". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

Rate-Limit & Loop Tripwire is a public reference for agent warnings, blocks, and current approval signals. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Plugin create". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Plugin inspizieren". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Plugin testen (execute)". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Plugin testen (smoke)". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Plugin aktivieren". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Plugin-Code read". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Plugin updaten". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Plugin rollback". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Plugin Permissions". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Plugin deinstallieren". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Managed Storage". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Plugin Export/Import". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Plugin-Liste". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Plugin deaktivieren". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Version-Code read". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Managed Storage read". It keeps purpose, scope, user control, and sensitive data handling explicit.

Multi-File Content Scripts is a public reference for extensions, external MCP sources, and bounded tool execution. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Runtime-/MCP-Tool-Grants is a public reference for extensions, external MCP sources, and bounded tool execution. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

ContentScript Inventory is a public reference for extensions, external MCP sources, and bounded tool execution. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Smoke ApplyWrites & Activate". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Plugin Bundle Discovery". It keeps purpose, scope, user control, and sensitive data handling explicit.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Memory save". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Memory fetch". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "Memory vergessen". It separates target confirmation, intended state change, and verification after execution.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Memory Stats". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Domain Exclude Gate". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Recall Access Tracking". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Instructions Memory Hints". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "Memory Forget Filter". It separates target confirmation, intended state change, and verification after execution.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to carry out the bounded step "Task create". It separates target confirmation, intended state change, and verification after execution.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to start or monitor the bounded flow "Task-Liste". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to carry out the bounded step "Task triggern". It separates target confirmation, intended state change, and verification after execution.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to carry out the bounded step "Run-Output read". It separates target confirmation, intended state change, and verification after execution.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to start or monitor the bounded flow "Workspace-Dateien". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to handle the sensitive path "Secrets". It keeps purpose, scope, user control, and sensitive data handling explicit.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to start or monitor the bounded flow "Task deaktivieren". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to carry out the bounded step "Delete task". It separates target confirmation, intended state change, and verification after execution.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to start or monitor the bounded flow "Export/Import". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to carry out the bounded step "Task updaten". It separates target confirmation, intended state change, and verification after execution.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to start or monitor the bounded flow "Task aktivieren". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to start or monitor the bounded flow "Task Details". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to carry out the bounded step "Run-History". It separates target confirmation, intended state change, and verification after execution.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to handle the sensitive path "Secret auflisten". It keeps purpose, scope, user control, and sensitive data handling explicit.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to start or monitor the bounded flow "Workspace read". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

Interaktiver Agent blockt Auto-Dispatch is a public reference for prepared tasks, runs, and closed-loop automation steps. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to handle the sensitive path "Trigger Inputs File". It keeps purpose, scope, user control, and sensitive data handling explicit.

Chain-Persistenz-Validierung is a public reference for prepared tasks, runs, and closed-loop automation steps. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Scheduled-Session Tool Policy is a public reference for prepared tasks, runs, and closed-loop automation steps. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to carry out the bounded step "Run Output Tail Read". It separates target confirmation, intended state change, and verification after execution.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to start or monitor the bounded flow "Workspace List Metadata". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to start or monitor the bounded flow "Scheduled Tasks Bundle Discovery". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to check the state or evidence for "Notification senden". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to check the state or evidence for "Notifications read". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to check the state or evidence for "Unread Count". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to check the state or evidence for "Mark Read". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to handle the sensitive path "Permissions". It keeps purpose, scope, user control, and sensitive data handling explicit.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to check the state or evidence for "Notification Details". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to carry out the bounded step "Notification open". It separates target confirmation, intended state change, and verification after execution.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to carry out the bounded step "Notification dismissn". It separates target confirmation, intended state change, and verification after execution.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to carry out the bounded step "Notifications clearen". It separates target confirmation, intended state change, and verification after execution.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to check the state or evidence for "Notifications Bundle". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to check the state or evidence for "Notification Send HighImpact". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to check the state or evidence for "Notification Spoofing Guard". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

Notification Permission Defaults is a public reference for user artifacts, notices, and small work handoffs. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to check the state or evidence for "Structured OK for Delivery-Status". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to carry out the bounded step "Note save". It separates target confirmation, intended state change, and verification after execution.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to check the state or evidence for "Note search". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to check the state or evidence for "Notes auflisten". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to carry out the bounded step "Note delete". It separates target confirmation, intended state change, and verification after execution.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to carry out the bounded step "Create favorite". It separates target confirmation, intended state change, and verification after execution.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to check the state or evidence for "Favorites read". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to carry out the bounded step "Favorite open". It separates target confirmation, intended state change, and verification after execution.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to check the state or evidence for "Favorite delete". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to check the state or evidence for "Clipboard". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to carry out the bounded step "Clipboard HighImpact Prevalidation". It separates target confirmation, intended state change, and verification after execution.

Clipboard Global Provenance is a public reference for user artifacts, notices, and small work handoffs. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to check the state or evidence for "Bookmark-Folder Discovery". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to carry out the bounded step "Bookmark-Folder Create/Rename". It separates target confirmation, intended state change, and verification after execution.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to carry out the bounded step "Bookmark-Folder Delete Modes". It separates target confirmation, intended state change, and verification after execution.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to carry out the bounded step "Favorite Folder Assignment & Move". It separates target confirmation, intended state change, and verification after execution.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to check the state or evidence for "Bookmark Alias Compatibility". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

Bookmark-Folder Negative Validation is a public reference for user artifacts, notices, and small work handoffs. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to check the state or evidence for "Bookmark Policy Buckets". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "Mobile-Viewport". It separates target confirmation, intended state change, and verification after execution.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "Touch aktivieren". It separates target confirmation, intended state change, and verification after execution.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "User-Agent". It separates target confirmation, intended state change, and verification after execution.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "Viewport-Emulation aufheben". It separates target confirmation, intended state change, and verification after execution.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to check the state or evidence for "Device-Emulation Bundle". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to handle the sensitive path "Dark-Mode emulieren". It keeps purpose, scope, user control, and sensitive data handling explicit.

Media-Features akkumulieren is a public reference for browser environment, window state, and isolated execution surfaces. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to handle the sensitive path "Media-Emulation aufheben". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Server add". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Server auflisten". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Server updaten". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Server start". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Server stop". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Server remove". It keeps purpose, scope, user control, and sensitive data handling explicit.

Unknown ServerKey is a public reference for extensions, external MCP sources, and bounded tool execution. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Transport-Validation is a public reference for extensions, external MCP sources, and bounded tool execution. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "External-MCP Bundle". It keeps purpose, scope, user control, and sensitive data handling explicit.

Download start is a public reference for user artifacts, notices, and small work handoffs. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to handle the sensitive path "Downloads read". It keeps purpose, scope, user control, and sensitive data handling explicit.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to handle the sensitive path "Download pausieren". It keeps purpose, scope, user control, and sensitive data handling explicit.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to handle the sensitive path "Download-Ordner". It keeps purpose, scope, user control, and sensitive data handling explicit.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to handle the sensitive path "Download abbrechen". It keeps purpose, scope, user control, and sensitive data handling explicit.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to handle the sensitive path "Download retrien". It keeps purpose, scope, user control, and sensitive data handling explicit.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to handle the sensitive path "Downloads clearen". It keeps purpose, scope, user control, and sensitive data handling explicit.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to handle the sensitive path "Open download". It keeps purpose, scope, user control, and sensitive data handling explicit.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to handle the sensitive path "Download Bulk Pause/Resume/Cancel". It keeps purpose, scope, user control, and sensitive data handling explicit.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to handle the sensitive path "Download Preview Passive Formats". It keeps purpose, scope, user control, and sensitive data handling explicit.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to handle the sensitive path "Download Preview Safety Blocks". It keeps purpose, scope, user control, and sensitive data handling explicit.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to handle the sensitive path "Download Auto-Open Roundtrip". It keeps purpose, scope, user control, and sensitive data handling explicit.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to handle the sensitive path "Download Auto-Open Blocklist". It keeps purpose, scope, user control, and sensitive data handling explicit.

Download Tool Policy Buckets is a public reference for user artifacts, notices, and small work handoffs. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For user artifacts, notices, and small work handoffs, an agent uses this public tool to handle the sensitive path "Auto-Open Input Shape". It keeps purpose, scope, user control, and sensitive data handling explicit.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to check the state or evidence for "UI-State". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "Open settings". It separates target confirmation, intended state change, and verification after execution.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to check the state or evidence for "Window-Groesse". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to check the state or evidence for "Zoom". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to check the state or evidence for "Native Dialog inspizieren". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "Native Dialog dismissn". It separates target confirmation, intended state change, and verification after execution.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to handle the sensitive path "Native Dialog File set". It keeps purpose, scope, user control, and sensitive data handling explicit.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "Native Dialog confirm". It separates target confirmation, intended state change, and verification after execution.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "Window State". It separates target confirmation, intended state change, and verification after execution.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to check the state or evidence for "Monitor-Inventar & Cross-Monitor-Move". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to handle the sensitive path "Downloads-Panel". It keeps purpose, scope, user control, and sensitive data handling explicit.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "Settings close". It separates target confirmation, intended state change, and verification after execution.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "WebView Zoom Reset". It separates target confirmation, intended state change, and verification after execution.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "DevTools Default Mode". It separates target confirmation, intended state change, and verification after execution.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to check the state or evidence for "App-Shell-Recovery Bundle". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "PDF Zoom MCP Guard". It separates target confirmation, intended state change, and verification after execution.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "App Quit". It separates target confirmation, intended state change, and verification after execution.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to carry out the bounded step "Run Sequence". It separates target confirmation, intended state change, and verification after execution.

Sequence with error is a public reference for prepared tasks, runs, and closed-loop automation steps. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to check the state or evidence for "Tab-Transfer". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to start or monitor the bounded flow "Sequence Common Arg Hints". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

Sequence No-Target Strictness is a public reference for prepared tasks, runs, and closed-loop automation steps. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Sequence Synthetic-Arg Matrix is a public reference for prepared tasks, runs, and closed-loop automation steps. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Sequence Pipeline-Arg Cleanup is a public reference for prepared tasks, runs, and closed-loop automation steps. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

License-Gate is a public reference for edge cases, product boundaries, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

License-Status is a public reference for edge cases, product boundaries, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For edge cases, product boundaries, and robust failure paths, an agent uses this public tool to check the state or evidence for "Get Onboarding". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For edge cases, product boundaries, and robust failure paths, an agent uses this public tool to carry out the bounded step "Install Onboarding". It separates target confirmation, intended state change, and verification after execution.

Self-Onboarding Advisory is a public reference for edge cases, product boundaries, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Bootstrap-Clear-Regeln is a public reference for edge cases, product boundaries, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Session-Latch is a public reference for edge cases, product boundaries, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For edge cases, product boundaries, and robust failure paths, an agent uses this public tool to check the state or evidence for "Reference Docs Liste". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For edge cases, product boundaries, and robust failure paths, an agent uses this public tool to check the state or evidence for "Reference Doc Read Paging & NotFound". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For risk-bounded actions with visible user control, an agent uses this public tool to start or monitor the bounded flow "Discover-Modus". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For risk-bounded actions with visible user control, an agent uses this public tool to start or monitor the bounded flow "Activate-Modus". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For risk-bounded actions with visible user control, an agent uses this public tool to start or monitor the bounded flow "Hover-Peek". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For risk-bounded actions with visible user control, an agent uses this public tool to carry out the bounded step "Close with health check". It separates target confirmation, intended state change, and verification after execution.

Safety-Gate-Block is a public reference for risk-bounded actions with visible user control. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For risk-bounded actions with visible user control, an agent uses this public tool to start or monitor the bounded flow "AX-Enrichment aktiv". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

AX-NameSource Scoring is a public reference for risk-bounded actions with visible user control. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Icon-Wrapper Name Walk is a public reference for risk-bounded actions with visible user control. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Read-Navigation Klasse is a public reference for risk-bounded actions with visible user control. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Deutsche Surface Labels is a public reference for risk-bounded actions with visible user control. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Surface Hard-Deny Rollen is a public reference for risk-bounded actions with visible user control. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

AX Negative Signal is a public reference for risk-bounded actions with visible user control. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

OPEN_NAVIGATION action category is a public reference for risk-bounded actions with visible user control. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For risk-bounded actions with visible user control, an agent uses this public tool to start or monitor the bounded flow "Surface-Explorer Bundle". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to check the state or evidence for "Goal create". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to check the state or evidence for "Goal abfragen". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to check the state or evidence for "Goal annotieren". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to check the state or evidence for "Goal close". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

Lease-Ablauf is a public reference for prepared tasks, runs, and closed-loop automation steps. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Preconditions is a public reference for prepared tasks, runs, and closed-loop automation steps. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to check the state or evidence for "goalContext in Perceive". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to carry out the bounded step "Goal Create Active Binding". It separates target confirmation, intended state change, and verification after execution.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to carry out the bounded step "Eval execute". It separates target confirmation, intended state change, and verification after execution.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to carry out the bounded step "Eval with DOM mutation". It separates target confirmation, intended state change, and verification after execution.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to carry out the bounded step "CDP Passthrough". It separates target confirmation, intended state change, and verification after execution.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to carry out the bounded step "Create Dump". It separates target confirmation, intended state change, and verification after execution.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to carry out the bounded step "Hardware Diagnostics". It separates target confirmation, intended state change, and verification after execution.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to handle the sensitive path "Media Status". It keeps purpose, scope, user control, and sensitive data handling explicit.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to check the state or evidence for "Traces read". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to carry out the bounded step "Create Dump ArtifactDir". It separates target confirmation, intended state change, and verification after execution.

For page perception, diagnostics, and verifiable read signals, an agent uses this public tool to carry out the bounded step "CDP Structured Result". It separates target confirmation, intended state change, and verification after execution.

For visual evidence, screenshots, and reviewable differences, an agent uses this public tool to check the state or evidence for "Stream-URL erzeugen". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For visual evidence, screenshots, and reviewable differences, an agent uses this public tool to check the state or evidence for "Screenshot-Qualitaet". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For visual evidence, screenshots, and reviewable differences, an agent uses this public tool to check the state or evidence for "Screenshot with resize". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For visual evidence, screenshots, and reviewable differences, an agent uses this public tool to check the state or evidence for "App-Screenshot". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

Screenshot Task-Isolation is a public reference for visual evidence, screenshots, and reviewable differences. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Screenshot ResponseMode Defaults is a public reference for visual evidence, screenshots, and reviewable differences. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For visual evidence, screenshots, and reviewable differences, an agent uses this public tool to check the state or evidence for "Screenshot Resource-URI". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

Screenshot Auto Mode is a public reference for visual evidence, screenshots, and reviewable differences. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Screenshot Dedupe is a public reference for visual evidence, screenshots, and reviewable differences. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For visual evidence, screenshots, and reviewable differences, an agent uses this public tool to carry out the bounded step "Screenshot Budget Gate". It separates target confirmation, intended state change, and verification after execution.

For visual evidence, screenshots, and reviewable differences, an agent uses this public tool to check the state or evidence for "App Screenshot Active-Target Guard". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For visual evidence, screenshots, and reviewable differences, an agent uses this public tool to check the state or evidence for "App Screenshot Unified Fields". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For visual evidence, screenshots, and reviewable differences, an agent uses this public tool to carry out the bounded step "Stream Preview Claim Gate". It separates target confirmation, intended state change, and verification after execution.

Stream Preview Query Validation is a public reference for visual evidence, screenshots, and reviewable differences. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For visual evidence, screenshots, and reviewable differences, an agent uses this public tool to check the state or evidence for "Full-Page-Screenshot". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "Platform Seed". It separates target confirmation, intended state change, and verification after execution.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Platform Get". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Platform List". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to check the state or evidence for "Sandbox-Kontext". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "Sandbox resolve". It separates target confirmation, intended state change, and verification after execution.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to check the state or evidence for "Tab in Sandbox". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "Sandbox create". It separates target confirmation, intended state change, and verification after execution.

Sandbox create (max) is a public reference for browser environment, window state, and isolated execution surfaces. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "Sandbox umbenennen". It separates target confirmation, intended state change, and verification after execution.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "Sandbox pausieren". It separates target confirmation, intended state change, and verification after execution.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "Delete sandbox". It separates target confirmation, intended state change, and verification after execution.

Letzte Sandbox schuetzen is a public reference for browser environment, window state, and isolated execution surfaces. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to check the state or evidence for "Sandbox CRUD Bundle Discovery". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For risk-bounded actions with visible user control, an agent uses this public tool to handle the sensitive path "Permissions read". It keeps purpose, scope, user control, and sensitive data handling explicit.

For risk-bounded actions with visible user control, an agent uses this public tool to handle the sensitive path "Permission set". It keeps purpose, scope, user control, and sensitive data handling explicit.

For risk-bounded actions with visible user control, an agent uses this public tool to handle the sensitive path "Media-Device Inventory Merge". It keeps purpose, scope, user control, and sensitive data handling explicit.

For risk-bounded actions with visible user control, an agent uses this public tool to handle the sensitive path "Preferred Device Web-ID Validation". It keeps purpose, scope, user control, and sensitive data handling explicit.

For connection setup, tool discovery, and transport readiness, an agent uses this public tool to carry out the bounded step "Restore-Prompt Status". It separates target confirmation, intended state change, and verification after execution.

For connection setup, tool discovery, and transport readiness, an agent uses this public tool to carry out the bounded step "Restore-Prompt beantworten". It separates target confirmation, intended state change, and verification after execution.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "Open DevTools". It separates target confirmation, intended state change, and verification after execution.

For browser environment, window state, and isolated execution surfaces, an agent uses this public tool to carry out the bounded step "Panel wechseln". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to check the state or evidence for "Layout Metrics". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For targeted browser interaction with a prechecked target, an agent uses this public tool to check the state or evidence for "Active Element Deep". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Scroll By". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Scroll To". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Scroll Smart". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Scroll Element". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Input Wheel". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Input Drag". It separates target confirmation, intended state change, and verification after execution.

For targeted browser interaction with a prechecked target, an agent uses this public tool to carry out the bounded step "Humanized Drag". It separates target confirmation, intended state change, and verification after execution.

For embedded agent runtimes and runtime communication, an agent uses this public tool to handle the sensitive path "Plugin Icons". It keeps purpose, scope, user control, and sensitive data handling explicit.

For embedded agent runtimes and runtime communication, an agent uses this public tool to handle the sensitive path "Plugin Security Log". It keeps purpose, scope, user control, and sensitive data handling explicit.

For embedded agent runtimes and runtime communication, an agent uses this public tool to handle the sensitive path "Plugin Logs". It keeps purpose, scope, user control, and sensitive data handling explicit.

For embedded agent runtimes and runtime communication, an agent uses this public tool to handle the sensitive path "Plugin Grant Active Tab". It keeps purpose, scope, user control, and sensitive data handling explicit.

For embedded agent runtimes and runtime communication, an agent uses this public tool to handle the sensitive path "Plugin CSS Reset". It keeps purpose, scope, user control, and sensitive data handling explicit.

For embedded agent runtimes and runtime communication, an agent uses this public tool to handle the sensitive path "Plugin Inject Reset". It keeps purpose, scope, user control, and sensitive data handling explicit.

UI Runtime-Port Sticky Handover is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Overlay Runtime-Port Snapshot is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Page Runtime-Port Snapshot is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Browser Tabs Runtime-Port is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For embedded agent runtimes and runtime communication, an agent uses this public tool to handle the sensitive path "Runtime Message Next Targets". It keeps purpose, scope, user control, and sensitive data handling explicit.

For embedded agent runtimes and runtime communication, an agent uses this public tool to handle the sensitive path "Runtime Message Known Sessions". It keeps purpose, scope, user control, and sensitive data handling explicit.

Browser Windows Runtime-Port is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For embedded agent runtimes and runtime communication, an agent uses this public tool to handle the sensitive path "Runtime Broadcast Port". It keeps purpose, scope, user control, and sensitive data handling explicit.

For embedded agent runtimes and runtime communication, an agent uses this public tool to handle the sensitive path "Runtime Worker Port". It keeps purpose, scope, user control, and sensitive data handling explicit.

Host-Peer Session Binding is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Max Sessions Per Profile is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Runtime Port Name Validation is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

EventDriven Idle Keepalive is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

UI Command & Context Menu Ports is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Port Query Commands is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For embedded agent runtimes and runtime communication, an agent uses this public tool to handle the sensitive path "Runtime Session Handles". It keeps purpose, scope, user control, and sensitive data handling explicit.

For embedded agent runtimes and runtime communication, an agent uses this public tool to handle the sensitive path "DocumentId Message Targeting". It keeps purpose, scope, user control, and sensitive data handling explicit.

For embedded agent runtimes and runtime communication, an agent uses this public tool to handle the sensitive path "Dynamic Content Scripts". It keeps purpose, scope, user control, and sensitive data handling explicit.

For embedded agent runtimes and runtime communication, an agent uses this public tool to handle the sensitive path "Runtime Permissions Self". It keeps purpose, scope, user control, and sensitive data handling explicit.

For embedded agent runtimes and runtime communication, an agent uses this public tool to handle the sensitive path "Subframe ElementHandle Adapter". It keeps purpose, scope, user control, and sensitive data handling explicit.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to handle the sensitive path "Variable set". It keeps purpose, scope, user control, and sensitive data handling explicit.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to handle the sensitive path "Variable read". It keeps purpose, scope, user control, and sensitive data handling explicit.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to handle the sensitive path "Variables auflisten". It keeps purpose, scope, user control, and sensitive data handling explicit.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to handle the sensitive path "Delete variable". It keeps purpose, scope, user control, and sensitive data handling explicit.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to handle the sensitive path "Templates". It keeps purpose, scope, user control, and sensitive data handling explicit.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to handle the sensitive path "Run abbrechen". It keeps purpose, scope, user control, and sensitive data handling explicit.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to handle the sensitive path "Active Runs". It keeps purpose, scope, user control, and sensitive data handling explicit.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to handle the sensitive path "Workspace auflisten". It keeps purpose, scope, user control, and sensitive data handling explicit.

For prepared tasks, runs, and closed-loop automation steps, an agent uses this public tool to handle the sensitive path "Workspace-Ordner". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Tool discovery with schema". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "External Tool Call". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Server Import". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Server Logs". It keeps purpose, scope, user control, and sensitive data handling explicit.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "Add candidate". It separates target confirmation, intended state change, and verification after execution.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Memory Stats". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "EVM Instructions Contract". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to carry out the bounded step "EVM LCJ Stats". It separates target confirmation, intended state change, and verification after execution.

For knowledge signals, learning candidates, and explainable orientation, an agent uses this public tool to check the state or evidence for "Explain PKS Level". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For browser identity, fingerprint surface, and profile impact, an agent uses this public tool to handle the sensitive path "Identity read". It keeps purpose, scope, user control, and sensitive data handling explicit.

For browser identity, fingerprint surface, and profile impact, an agent uses this public tool to handle the sensitive path "Identity-Presets". It keeps purpose, scope, user control, and sensitive data handling explicit.

For browser identity, fingerprint surface, and profile impact, an agent uses this public tool to handle the sensitive path "Identity set". It keeps purpose, scope, user control, and sensitive data handling explicit.

Identity + Navigation is a public reference for browser identity, fingerprint surface, and profile impact. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For browser identity, fingerprint surface, and profile impact, an agent uses this public tool to handle the sensitive path "Custom-UA". It keeps purpose, scope, user control, and sensitive data handling explicit.

For browser identity, fingerprint surface, and profile impact, an agent uses this public tool to handle the sensitive path "Partial Update". It keeps purpose, scope, user control, and sensitive data handling explicit.

For browser identity, fingerprint surface, and profile impact, an agent uses this public tool to handle the sensitive path "Idempotenz". It keeps purpose, scope, user control, and sensitive data handling explicit.

For browser identity, fingerprint surface, and profile impact, an agent uses this public tool to handle the sensitive path "Validation". It keeps purpose, scope, user control, and sensitive data handling explicit.

For browser identity, fingerprint surface, and profile impact, an agent uses this public tool to handle the sensitive path "Fingerprint Bundle Discovery". It keeps purpose, scope, user control, and sensitive data handling explicit.

For browser identity, fingerprint surface, and profile impact, an agent uses this public tool to handle the sensitive path "Fingerprint Effective Hierarchy". It keeps purpose, scope, user control, and sensitive data handling explicit.

For browser identity, fingerprint surface, and profile impact, an agent uses this public tool to handle the sensitive path "Fingerprint Global Persistence". It keeps purpose, scope, user control, and sensitive data handling explicit.

For browser identity, fingerprint surface, and profile impact, an agent uses this public tool to handle the sensitive path "Fingerprint Sandbox Override". It keeps purpose, scope, user control, and sensitive data handling explicit.

For browser identity, fingerprint surface, and profile impact, an agent uses this public tool to handle the sensitive path "Fingerprint Tab Override". It keeps purpose, scope, user control, and sensitive data handling explicit.

Fingerprint Technique Coverage is a public reference for browser identity, fingerprint surface, and profile impact. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Fingerprint Validation & Policy is a public reference for browser identity, fingerprint surface, and profile impact. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For browser identity, fingerprint surface, and profile impact, an agent uses this public tool to handle the sensitive path "Identity-Management Bundle". It keeps purpose, scope, user control, and sensitive data handling explicit.

For browser identity, fingerprint surface, and profile impact, an agent uses this public tool to handle the sensitive path "Fingerprint Bundle Metadata Consistency". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Stdio-Transport Handshake". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "HTTP-Transport Handshake". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "SSE-Transport Handshake (Legacy)". It keeps purpose, scope, user control, and sensitive data handling explicit.

Multi-Server parallel is a public reference for extensions, external MCP sources, and bounded tool execution. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Hub + Nova-Tools im selben Flow". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Inventory Refresh nach Reconnect". It keeps purpose, scope, user control, and sensitive data handling explicit.

Crash/Disconnect Recovery is a public reference for extensions, external MCP sources, and bounded tool execution. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Server Presets is a public reference for extensions, external MCP sources, and bounded tool execution. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Server-Lifecycle ohne Orphan is a public reference for extensions, external MCP sources, and bounded tool execution. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Discovery Auto-Registration". It keeps purpose, scope, user control, and sensitive data handling explicit.

For extensions, external MCP sources, and bounded tool execution, an agent uses this public tool to handle the sensitive path "Discovery Trust-Gate". It keeps purpose, scope, user control, and sensitive data handling explicit.

Claude Sidecar start is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Codex Sidecar start is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Gemini Sidecar start is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Sidecar-Discovery is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Sidecar MCP-Claim is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Sidecar Session-Resume is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Emergency Stop is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Crash-Loop Detection is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Sidecar-Repair is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Raw-Log Forensics is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Conversation Titles is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Multi-Agent parallel is a public reference for embedded agent runtimes and runtime communication. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Token-Generierung is a public reference for MCP configuration, token sync, and recovery. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Auto-Sync Claude is a public reference for MCP configuration, token sync, and recovery. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Auto-Sync Codex is a public reference for MCP configuration, token sync, and recovery. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Manueller Sync is a public reference for MCP configuration, token sync, and recovery. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Port-Wechsel is a public reference for MCP configuration, token sync, and recovery. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Stale-Token Recovery is a public reference for MCP configuration, token sync, and recovery. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Startup Config Scan is a public reference for MCP configuration, token sync, and recovery. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For MCP configuration, token sync, and recovery, an agent uses this public tool to handle the sensitive path "Codex Proxy Backend Session Refresh". It keeps purpose, scope, user control, and sensitive data handling explicit.

Discovery File Shape is a public reference for MCP configuration, token sync, and recovery. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Codex Proxy Header Sync is a public reference for MCP configuration, token sync, and recovery. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For visual evidence, screenshots, and reviewable differences, an agent uses this public tool to carry out the bounded step "Highlight-Selector". It separates target confirmation, intended state change, and verification after execution.

For visual evidence, screenshots, and reviewable differences, an agent uses this public tool to carry out the bounded step "Scroll-to-Selector". It separates target confirmation, intended state change, and verification after execution.

Shadow-DOM Highlight is a public reference for visual evidence, screenshots, and reviewable differences. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For visual evidence, screenshots, and reviewable differences, an agent uses this public tool to start or monitor the bounded flow "Crawl-Verify Highlight". It names the scope, progress signal, stopping condition, and result check that keep the run bounded.

For visual evidence, screenshots, and reviewable differences, an agent uses this public tool to check the state or evidence for "Perceive + Screenshot". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For visual evidence, screenshots, and reviewable differences, an agent uses this public tool to carry out the bounded step "Navigation + Screenshot". It separates target confirmation, intended state change, and verification after execution.

For visual evidence, screenshots, and reviewable differences, an agent uses this public tool to carry out the bounded step "Click + Screenshot". It separates target confirmation, intended state change, and verification after execution.

Crawl -> Learn Loop is a public reference for edge cases, product boundaries, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Login-Komplett-Flow is a public reference for edge cases, product boundaries, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Plugin bauen + nutzen is a public reference for edge cases, product boundaries, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Task-Memory E2E is a public reference for edge cases, product boundaries, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Claim + Navigation + Release is a public reference for edge cases, product boundaries, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Surface Explorer + PKS is a public reference for edge cases, product boundaries, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Scheduled Task + Workspace is a public reference for edge cases, product boundaries, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Multi-Tab Workflow is a public reference for edge cases, product boundaries, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Notification + Agent-Aktion is a public reference for edge cases, product boundaries, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Proxy + Navigation is a public reference for edge cases, product boundaries, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Download + Workspace is a public reference for edge cases, product boundaries, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Emulation + Perceive is a public reference for edge cases, product boundaries, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Goal + Sequence is a public reference for edge cases, product boundaries, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Vault + Plugin Login is a public reference for edge cases, product boundaries, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Invalid targetId is a public reference for parameter validation, alias conflicts, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Unbekannte Property is a public reference for parameter validation, alias conflicts, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For parameter validation, alias conflicts, and robust failure paths, an agent uses this public tool to check the state or evidence for "String statt Number". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For parameter validation, alias conflicts, and robust failure paths, an agent uses this public tool to carry out the bounded step "Out-of-Range Integer". It separates target confirmation, intended state change, and verification after execution.

Claim-Session-Mismatch is a public reference for parameter validation, alias conflicts, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For parameter validation, alias conflicts, and robust failure paths, an agent uses this public tool to carry out the bounded step "Release ohne Claim". It separates target confirmation, intended state change, and verification after execution.

Sequence step error is a public reference for parameter validation, alias conflicts, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Tab close with pending state is a public reference for parameter validation, alias conflicts, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For parameter validation, alias conflicts, and robust failure paths, an agent uses this public tool to carry out the bounded step "Doppelter Click-Guard". It separates target confirmation, intended state change, and verification after execution.

For parameter validation, alias conflicts, and robust failure paths, an agent uses this public tool to carry out the bounded step "Type with hydration drift". It separates target confirmation, intended state change, and verification after execution.

PKS Shadow-Persistence is a public reference for parameter validation, alias conflicts, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For parameter validation, alias conflicts, and robust failure paths, an agent uses this public tool to carry out the bounded step "Concurrent Route Conflict". It separates target confirmation, intended state change, and verification after execution.

Non-Object Arguments Matrix is a public reference for parameter validation, alias conflicts, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Unknown Properties vor Gates is a public reference for parameter validation, alias conflicts, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For parameter validation, alias conflicts, and robust failure paths, an agent uses this public tool to carry out the bounded step "Sensitive Eval HighImpact". It separates target confirmation, intended state change, and verification after execution.

For parameter validation, alias conflicts, and robust failure paths, an agent uses this public tool to check the state or evidence for "Non-Web Navigation HighImpact". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

For parameter validation, alias conflicts, and robust failure paths, an agent uses this public tool to check the state or evidence for "Tool-Alias Dispatch Matrix". It explains which current signal can be used as evidence and when a follow-up step needs a fresh target check.

Param-Alias Canonical Wins is a public reference for parameter validation, alias conflicts, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Common Arg Alias Normalization is a public reference for parameter validation, alias conflicts, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Param-Alias Scope Negative is a public reference for parameter validation, alias conflicts, and robust failure paths. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Bundle laden". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Cookie-List Metadaten". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Cookie list with values". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Cookie-List Filter". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Cookie-List Pagination". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Cookie-Set dryRun". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Cookie-Set create". It keeps purpose, scope, user control, and sensitive data handling explicit.

Cookie-Set erset is a public reference for sensitive browser, identity, or session data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Cookie-Set Domain-Reject". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Cookie-Set Public-Suffix". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Cookie-Set SameSite-Reject". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Cookie-Delete per cookieId". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Cookie-Delete per Name". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Cookie-Delete dryRun". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Cookie-Clear HighImpact (profil-weit)". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Storage-Inspect Keys". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Storage-Inspect Values". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Storage-Set". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Storage-Delete". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Cache-Clear". It keeps purpose, scope, user control, and sensitive data handling explicit.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Cache-Clear Unknown Type". It keeps purpose, scope, user control, and sensitive data handling explicit.

Scope-Sichtbarkeit is a public reference for sensitive browser, identity, or session data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Sandbox-Isolation is a public reference for sensitive browser, identity, or session data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Permission-Deny is a public reference for sensitive browser, identity, or session data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Permission-Session-Grant is a public reference for sensitive browser, identity, or session data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Cookie-Inspector Service-Paritaet". It keeps purpose, scope, user control, and sensitive data handling explicit.

Cookie-Inspector Editor Controls is a public reference for sensitive browser, identity, or session data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

HttpOnly Reveal/Copy Gate is a public reference for sensitive browser, identity, or session data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Sandbox Account/Badge Refresh is a public reference for sensitive browser, identity, or session data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

NovaPalette Cookie-Inspector Theme is a public reference for sensitive browser, identity, or session data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

For sensitive browser, identity, or session data, an agent uses this public tool to handle the sensitive path "Cookie-Values Scope Required". It keeps purpose, scope, user control, and sensitive data handling explicit.

SiteData Agent-Scoped Grants is a public reference for sensitive browser, identity, or session data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.

Secret-Read Policy Upgrade is a public reference for sensitive browser, identity, or session data. It names the signal, policy, or flow an agent should understand before choosing a concrete tool.