Release Judge — Security agent for Claude Code
Reviews the full evidence bundle for a release candidate (gate results, test and security summaries, triage outputs, requirement analyses) and writes an advisory go / no-go recommendation with reasons.
How to install Release Judge
Installs to ~/.claude/agents/dbisina-qapro-release-judge.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/dbisina/QAPRO/HEAD/.claude/agents/release-judge.md -o ~/.claude/agents/dbisina-qapro-release-judge.md Restart Claude Code, or start a new session, for it to be picked up.
What Release Judge does
name: release-judge description: Reviews the full evidence bundle for a release candidate (gate results, test and security summaries, triage outputs, requirement analyses) and writes an advisory go / no-go recommendation with reasons. Opus, because it verifies what the other agents surfaced. tools: Read, Grep, Glob model: opus
You are the release judge. You verify the work of the other agents and give the humans who approve production a clear, honest recommendation.
Inputs
- The gat
Alternatives in Security
- Security Requirement Extraction — Transform threat analysis into actionable security requirements 31.9k ★
- Audit Deep — Deep single-subsystem audit that reasons about state over time (caches, upserts, migrations, concurrent scans) 335 ★
- Advisory Hunter — Phase 1 intelligence gathering agent that collects security advisories (CVE, GHSA, OSV) with adaptive time exp 125 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
AI Judge
Gatekeeper agent for structured rubric scoring of research, architecture, security, and planning outputs. Vali
QA Release Ops
ADLC Agent 4 — integration/regression tests, eval suite, blocking security gate on stage:qa PRs; drafts the Ga
Package Checker
Dependency auditor agent — validates package.json and lockfiles (versions, integrity, deprecated/security, bun
Security Triage
Independently evaluate a single security finding to decide whether it is a real, exploitable vulnerability or
Cofounder Agent
Strategy-layer assistant for a program's cofounder. Helps author and audit the six v1 strategy registries (pit
Security Requirements Reviewer
Read-only security architect's review of a project's requirement/spec set (requirements, ADRs, spec docs) for
Related Skills
Audio Compare
Advisory before/after audio judgment — measure → compare → judge two renders and weigh in on a DSP change with
Mem Lint
Heuristic contradiction lint across [ref]/[decision]/[tool] entries — flags polarity mismatches (enabled vs di
Knowgt
Generate concise, human-readable GitHub Trending briefings from WebFetch/WebSearch results, compare against pr