By tailrocks
High-signal pull request review workflow with specialized review agents, validation, and fix guidance
Adversarially stress-tests PR assumptions, decision dependencies, hidden constraints, edge cases, and unresolved tradeoffs.
High-signal PR reviewer for scoped project rules, real bugs, serious regressions, and important maintainability defects.
Finds behavior-preserving simplifications for recently changed code after correctness and safety review.
Checks changed comments, docs, examples, help text, and TODOs for accuracy, usefulness, and maintainability.
Inspects only the patch for obvious high-confidence compile, parse, type, import, reference, and logic failures.
Uses power tools
Uses Bash, Write, or Edit tools
No model invocation
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Executes directly as bash, bypassing the AI model
Executes directly as bash, bypassing the AI model
High-signal pull request review for Claude Code, Codex, and Amp.
This plugin packages a serious review workflow that ports the behavior of Anthropic's Code Review, Code Simplifier, and PR Review Toolkit plugins using original prompt text released under Apache-2.0. It is designed to find merge-blocking bugs, missed edge cases, weak tests, silent failures, security boundary mistakes, type invariant problems, misleading comments, and unnecessary complexity.
| Component | Purpose |
|---|---|
/review-crucible:code-review | Claude-style strict PR review with high-confidence filtering |
/review-crucible:review-pr | Claude Code slash command for a full PR or diff review |
/review-crucible:simplify | Claude Code command for conservative behavior-preserving simplification |
review-crucible skill | Source-neutral Codex/Claude skill for orchestrating review |
| Review agents | Focused Claude Code subagents for different failure modes |
| Codex custom agents | .codex/agents/*.toml reviewers for Codex subagent fan-out |
| Amp checks | .agents/checks/*.md criteria for Amp code review |
| Install snippets | Codex and Amp setup instructions |
Canonical review content lives under skills/review-crucible/references/.
Host files are thin wrappers or synced host copies:
reviewer-roles.md: reviewer roles and preserved source-plugin concepts.review-orchestration.md: preflight, pass selection, targeted activation,
host routing, and validation ownership.review-rubric.md: severity, confidence, validation, false-positive filters,
and output format.github-commenting.md: inline review comment discipline.model-policy.md: strongest verified model and highest effort policy.Reviewer roster, host frontmatter (model/effort/color, Codex reasoning effort,
Amp severity/tools), and plugin manifest metadata live in plugin.shared.json.
Wrappers under agents/, .codex/agents/, .agents/checks/, and both plugin
manifests are generated. Edit canonical references and plugin.shared.json,
then run a single command:
python3 scripts/regen.py
Regeneration also removes stale generated wrappers after role/check renames,
but only when the stale file still carries the Generated by scripts/regen.py
banner.
CI / pre-commit drift check (exits nonzero if any wrapper is out of sync or a stale generated wrapper remains):
python3 scripts/regen.py --check
Full local validation:
python3 scripts/validate.py
That runner checks generated wrappers, manifests, Codex agent TOML, Python scripts, shared metadata anchors and duplicate generated names, the Review Crucible skill validator when available, and worktree and staged whitespace-sensitive diffs.
| Agent | How Review Crucible is used | Subagent behavior |
|---|---|---|
| Claude Code | Native plugin commands and auto-discovered agents | Plugin agents run as subagents |
| Codex | Reusable skill, .codex-plugin/plugin.json, and .codex/agents/*.toml | Uses Codex subagents/custom agents in current Codex app/CLI; otherwise sequential passes |
| Amp | AGENTS.md snippet plus .agents/checks/*.md review checks | Review checks run as separate review subagents |
See docs/COMPATIBILITY.md for details and host limits. See docs/PARITY-AUDIT.md for source-plugin behavior mapping.
rust-best-practices when that plugin or skill
is installed.Install from the TailRocks marketplace:
/plugin marketplace add tailrocks/tailrocks-marketplace
/plugin install review-crucible@tailrocks-marketplace
For local development from a clone, add the clone path as a marketplace:
/plugin marketplace add /absolute/path/to/tailrocks-marketplace
/plugin install review-crucible@tailrocks-marketplace
/reload-plugins
This plugin does not ship its own marketplace manifest. Distribution is handled by the TailRocks marketplace repository.
Run:
npx claudepluginhub tailrocks/review-crucibleRust engineering skills for writing, reviewing, refactoring, and documenting idiomatic Rust code
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
Comprehensive PR review agents specializing in comments, tests, error handling, type design, code quality, and code simplification
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.