This repository also configures its own agents. See what case tells them →
Agent Claude Code
PR creation agent for /case. Drafts thorough PR descriptions from task file and verification evidence. Verifies all evidence gates before PR creation. Never implements or tests.
Agent Claude Code
Focused code implementation agent for /case. Writes fixes, runs unit tests, commits. Does not handle manual testing, evidence, or PRs.
Agent Claude Code
Read-only repo interviewer. Explores a target repo and asks the human targeted questions so ca onboard --interview can persist a correct evidence strategy, verification notes, and seed knowledge files.
Agent Claude Code
Post-run analysis agent for /case. Reads the progress log, identifies harness improvements, and proposes amendments for human review. Only repo learnings are applied directly.
Agent Claude Code
Code review agent for /case. Reads the diff against golden principles and structured test output. Produces findings that gate PR creation (critical) or inform via PR comments (warning/info). Never implements or tests.
Agent Claude Code
Read-only exploration agent. Runs before the implementer to surface relevant files, patterns, and constraints so the implementer starts with concrete context.
Agent Claude Code
Fresh-context verification agent for /case. Reads the diff, tests the specific fix with Playwright, creates evidence markers and screenshots. Never implements.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: