From agent-validator
Summarizes latest validator session with overview, gate pass/fail results, and detailed failures from check logs (linters, tests) and review JSONs (violations by file/line/issue). Useful for reviewing validation runs.
npx claudepluginhub codagent-ai/agent-validator --plugin agent-validatorThis skill is limited to using the following tools:
Show a detailed summary of the most recent validator session.
Diagnoses validator failures and explains behavior from runtime artifacts like .debug.log, .execution_state, console logs, check outputs, and config.yml. For 'why did validator fail' or 'diagnose logs' queries.
Fixes session protocol validation failures in GitHub Actions. Reads Job Summary from failed PR runs to parse NON_COMPLIANT verdicts, corrects session log structure per requirements, and commits fixes.
Fixes session protocol validation failures in GitHub Actions by reading Job Summary, analyzing session logs against SESSION-PROTOCOL.md, and applying targeted fixes. Use for PRs failing with NON_COMPLIANT verdicts or MUST requirements.
Share bugs, ideas, or general feedback.
Show a detailed summary of the most recent validator session.
agent-validate status 2>&1
The script parses the .debug.log for session-level data (run count, gate results, pass/fail status) and lists all log files with their paths and sizes.
For each gate marked FAIL in the Gate Results table, read the corresponding log files to extract failure details:
check:src:code-health): Read the matching check_*.log file. Check log formats vary by tool (linters, test runners, code health analyzers) — read the file and extract the relevant error/warning output.review:.:code-quality): Read the matching review_*.json file(s). These contain structured violation data with file, line, issue, priority, and status fields.Use the file paths from the "Log Files" section of the script output. Match gate IDs to file names: check:.:lint corresponds to check_._lint.*.log, review:.:code-quality corresponds to review_._code-quality_*.{log,json}.
Combine the script's session summary with the detailed failure information into a comprehensive report: