Validate agent delegation outputs and produce PASS or FAIL verdicts.
npx skills add https://github.com/etylsarin/opencastle --skill fast-review
Please help me install this Agent Skill. Skill: fast-review Source: https://github.com/etylsarin/opencastle/tree/main/src/orchestrator/skills/fast-review Command: npx skills add https://github.com/etylsarin/opencastle --skill fast-review
Fast Review prevents unchecked agent outputs from progressing in automated workflows by providing a lightweight, mandatory verification step that catches regressions, boundary violations, and semantic errors before acceptance. It reduces risk during overnight or long-running autonomous sessions by ensuring each delegation is validated quickly and consistently.
Run fast-review on the agent's output to obtain a PASS or FAIL verdict with structured issues, feedback, and recommendations for retry or escalation.
High-intent search queries and answers about installing and using this skill.
You can gate agent delegations by running a single-review verification step that evaluates outputs against acceptance criteria, producing structured PASS or FAIL verdicts to catch regressions, boundary violations, and semantic errors before acceptance.
When automated code review detects a failure, the system retries the same agent up to two times with actionable feedback. If failures persist, it escalates the output to a 3-reviewer panel review and records disputes for human resolution.
Yes, automated review gates are designed for overnight autonomous runs. They provide lightweight, consistent validation for routine code changes and low- to medium-risk delegations, reducing the risk of unchecked outputs progressing unattended.
Yes, the review gate integrates with deterministic checks by running after lint, test, build, secret scanning, and blast-radius checks, and it requires logging before proceeding to ensure comprehensive validation coverage.
You should escalate to a panel review after a sub-agent fails the validation gate twice. The system automatically retries the delegation up to two times on FAIL verdicts before triggering a 3-reviewer panel escalation for dispute resolution.
etylsarinScan code changes for secrets and run deterministic validation gates.
michael-conradVerify text, vision, and audio claims against evidence with PASS/FAIL/UNVERIFIED results.
fockusCoordinate parallel sub-agents and an automated judge to complete tasks with retries.
yamadashyAutomate iterative code review and fix cycles with Claude Code reviewer sub-agent.
