Repository navigation
Conversation
Requests whose ancestry, environment names or terminal session mark them as an AI coding tool carry the ai-agent label in the Cedar context, and the built-in policy asks for each of them instead of reusing the terminal's grant. `mairu lock --tool` marks a session; `mairu unlock --tool` always asks before removing a mark. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Dg17LE9aWqsb1VBUCVcKLM
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Dg17LE9aWqsb1VBUCVcKLM
sorah
changed the title
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Implements docs/specs/010-tool-detection.md.
ai-agentfrom its ancestry (comm,exe,argv[0..1]basenames, pidfd-pinned walk up to 64), environment variable names the client reports (names only), or session markscontext.tools/context.tool_matchesat both gates;builtin-humanexcludesai-agent, andbuiltin-ai-agent-per-requestasks for each such request (or a runningmairu execlease idle at most 1 h)mairu lock --tool <label>marks the terminal session;mairu unlock --toolalways asks through the approve-policy actionfrom:line,doctorand the decision log show labelsCODEX_THREAD_ID,CURSOR_AGENT,QWEN_CODE,OPENCODE,GOOSE_TERMINAL,AMP_CURRENT_THREAD_ID,COPILOT_CLI,CRUSH)Verified in Docker (35 checks): requests under fake
claude/codexprocesses or withCLAUDECODEask every time while a human terminal asks once;mairu execleases; marked panes;unlock --tool.🤖 Generated with Claude Code
https://claude.ai/code/session_01Dg17LE9aWqsb1VBUCVcKLM