mayankmankhand/llm-peer-review
Multi-model AI debate for your entire project lifecycle. Plans, specs, research, and code. Free and open source.
ARCHETYPE
Orchestrator
A bench of specialised subagents. The main agent mostly delegates.
Copy this rig
# review before running: this installs third-party code
$ npx degit mayankmankhand/llm-peer-review/.claude ./rig-llm-peer-review # inspect, then merge into .claude/ MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.
$ npx degit mayankmankhand/llm-peer-review/.claude/skills/audit-html .claude/skills/audit-html $ npx degit mayankmankhand/llm-peer-review/.claude/skills/design-rules .claude/skills/design-rules $ npx degit mayankmankhand/llm-peer-review/.claude/skills/dispatch-contract .claude/skills/dispatch-contract $ npx degit mayankmankhand/llm-peer-review/.claude/skills/error-analysis .claude/skills/error-analysis $ npx degit mayankmankhand/llm-peer-review/.claude/skills/html-viewing .claude/skills/html-viewing $ npx degit mayankmankhand/llm-peer-review/.claude/skills/learning-opportunity .claude/skills/learning-opportunity $ npx degit mayankmankhand/llm-peer-review/.claude/skills/playground .claude/skills/playground $ npx degit mayankmankhand/llm-peer-review/.claude/skills/project-context .claude/skills/project-context $ npx degit mayankmankhand/llm-peer-review/.claude/skills/review-browser-criteria .claude/skills/review-browser-criteria $ npx degit mayankmankhand/llm-peer-review/.claude/skills/review-browser .claude/skills/review-browser $ npx degit mayankmankhand/llm-peer-review/.claude/skills/review-code-criteria .claude/skills/review-code-criteria $ npx degit mayankmankhand/llm-peer-review/.claude/skills/review-code .claude/skills/review-code $ npx degit mayankmankhand/llm-peer-review/.claude/skills/review-commands-criteria .claude/skills/review-commands-criteria $ npx degit mayankmankhand/llm-peer-review/.claude/skills/review-commands .claude/skills/review-commands $ npx degit mayankmankhand/llm-peer-review/.claude/skills/review-copy-criteria .claude/skills/review-copy-criteria $ npx degit mayankmankhand/llm-peer-review/.claude/skills/review-copy .claude/skills/review-copy $ npx degit mayankmankhand/llm-peer-review/.claude/skills/review-deps-criteria .claude/skills/review-deps-criteria $ npx degit mayankmankhand/llm-peer-review/.claude/skills/review-deps .claude/skills/review-deps $ npx degit mayankmankhand/llm-peer-review/.claude/skills/review-full .claude/skills/review-full $ npx degit mayankmankhand/llm-peer-review/.claude/skills/review-plan-criteria .claude/skills/review-plan-criteria $ npx degit mayankmankhand/llm-peer-review/.claude/skills/review-plan .claude/skills/review-plan $ npx degit mayankmankhand/llm-peer-review/.claude/skills/review-security-criteria .claude/skills/review-security-criteria $ npx degit mayankmankhand/llm-peer-review/.claude/skills/review-security .claude/skills/review-security $ npx degit mayankmankhand/llm-peer-review/.claude/skills/review-ux-criteria .claude/skills/review-ux-criteria $ npx degit mayankmankhand/llm-peer-review/.claude/skills/review-ux .claude/skills/review-ux $ npx degit mayankmankhand/llm-peer-review/.claude/skills/security-audit .claude/skills/security-audit $ npx degit mayankmankhand/llm-peer-review/.claude/skills/setup .claude/skills/setup $ npx degit mayankmankhand/llm-peer-review/.claude/skills/upgrade .claude/skills/upgrade
$ curl -fsSL --create-dirs -o .claude/agents/audit-skeptic.md https://raw.githubusercontent.com/mayankmankhand/llm-peer-review/main/.claude/agents/audit-skeptic.md $ curl -fsSL --create-dirs -o .claude/agents/correction-extractor.md https://raw.githubusercontent.com/mayankmankhand/llm-peer-review/main/.claude/agents/correction-extractor.md $ curl -fsSL --create-dirs -o .claude/agents/design-comparer.md https://raw.githubusercontent.com/mayankmankhand/llm-peer-review/main/.claude/agents/design-comparer.md $ curl -fsSL --create-dirs -o .claude/agents/design-critic.md https://raw.githubusercontent.com/mayankmankhand/llm-peer-review/main/.claude/agents/design-critic.md $ curl -fsSL --create-dirs -o .claude/agents/fix-verifier.md https://raw.githubusercontent.com/mayankmankhand/llm-peer-review/main/.claude/agents/fix-verifier.md $ curl -fsSL --create-dirs -o .claude/agents/index-mapper.md https://raw.githubusercontent.com/mayankmankhand/llm-peer-review/main/.claude/agents/index-mapper.md $ curl -fsSL --create-dirs -o .claude/agents/plan-critic.md https://raw.githubusercontent.com/mayankmankhand/llm-peer-review/main/.claude/agents/plan-critic.md $ curl -fsSL --create-dirs -o .claude/agents/review-browser-finder.md https://raw.githubusercontent.com/mayankmankhand/llm-peer-review/main/.claude/agents/review-browser-finder.md $ curl -fsSL --create-dirs -o .claude/agents/review-code-finder.md https://raw.githubusercontent.com/mayankmankhand/llm-peer-review/main/.claude/agents/review-code-finder.md $ curl -fsSL --create-dirs -o .claude/agents/review-commands-finder.md https://raw.githubusercontent.com/mayankmankhand/llm-peer-review/main/.claude/agents/review-commands-finder.md $ curl -fsSL --create-dirs -o .claude/agents/review-copy-finder.md https://raw.githubusercontent.com/mayankmankhand/llm-peer-review/main/.claude/agents/review-copy-finder.md $ curl -fsSL --create-dirs -o .claude/agents/review-deps-finder.md https://raw.githubusercontent.com/mayankmankhand/llm-peer-review/main/.claude/agents/review-deps-finder.md $ curl -fsSL --create-dirs -o .claude/agents/review-finder.md https://raw.githubusercontent.com/mayankmankhand/llm-peer-review/main/.claude/agents/review-finder.md $ curl -fsSL --create-dirs -o .claude/agents/review-plan-finder.md https://raw.githubusercontent.com/mayankmankhand/llm-peer-review/main/.claude/agents/review-plan-finder.md $ curl -fsSL --create-dirs -o .claude/agents/review-security-finder.md https://raw.githubusercontent.com/mayankmankhand/llm-peer-review/main/.claude/agents/review-security-finder.md $ curl -fsSL --create-dirs -o .claude/agents/review-ux-finder.md https://raw.githubusercontent.com/mayankmankhand/llm-peer-review/main/.claude/agents/review-ux-finder.md
This rig commits no guardrails. Here is the community baseline instead — the deny/ask rules most often found across all 16.2k rigs:
{
"permissions": {
"deny": [
"Read(./.env)",
"Read(./.env.*)",
"Read(**/.env)",
"Read(~/.ssh/**)",
"Bash(git push --force:*)",
"Read(**/*.pem)",
"Read(.env)",
"Read(~/.aws/**)",
"Read(**/*.key)",
"Read(**/.env.*)",
"Bash(git push --force*)",
"Bash(rm -rf:*)",
"Bash(rm -rf *)",
"Bash(sudo *)",
"Bash(sudo:*)",
"Bash(rm -rf /)",
"Bash(git push -f:*)",
"Read(.env.*)",
"Bash(git reset --hard*)",
"Bash(rm -rf /*)"
],
"ask": [
"Bash(git push:*)",
"Bash(git push *)",
"Bash(gh pr merge *)",
"Bash(git commit:*)",
"Bash(rm *)",
"Bash(gh pr merge:*)",
"Bash(gh pr create *)",
"Bash(git reset --hard:*)",
"Bash(npm publish:*)",
"Bash(git rebase *)"
]
}
} Skills (28)
audit-htmldesign-rulesdispatch-contracterror-analysishtml-viewinglearning-opportunityplaygroundproject-contextreview-browser-criteriareview-browserreview-code-criteriareview-codereview-commands-criteriareview-commandsreview-copy-criteriareview-copyreview-deps-criteriareview-depsreview-fullreview-plan-criteriareview-planreview-security-criteriareview-securityreview-ux-criteriareview-uxsecurity-auditsetupupgrade
Subagents (16)
| audit-skeptic model: inherit | M2 audit skeptic. A fresh-context judge that receives findings as verbatim bytes with their receipt output and tries to refute each one - a tier-2 shard over Warns and Suggests, or one of the three ti |
| correction-extractor model: haiku | Correction-ledger worker for /document's capture stage. Reads pre-filtered transcript candidates cold and reports which were genuine human interventions, with a plain-language open code for each. Read |
| design-comparer model: inherit | Side-by-side design judge for /execute design steps (M15). Receives two screenshot paths labeled A and B and the fixed comparer prompt from design-rules.md, and returns which one is closer to a top-st |
| design-critic model: inherit | Design critic worker for /execute design steps (M15). Receives one screenshot path and the fixed critic prompt from design-rules.md, judges the design against a top-studio bar, and returns the biggest |
| fix-verifier model: inherit | M3 judgment verifier. A fresh-context judge that receives a shard of judgment findings with their original text, file:line, and the diff of the fixes, and returns one line per ID - "RN: FIXED" or "RN: |
| index-mapper model: sonnet | Codebase-chunk analysis worker for /index Step 3. Reads the files in its assigned chunk and emits structured module blocks per the dispatching prompt's exact format. Read-only. |
| plan-critic model: inherit | Plan critic for /create-plan. A fresh-context judge that reads one plan file and the exploration's closing summary, judges the plan against a staff-level bar, and returns its biggest gaps, or the lite |
| review-browser-finder model: opus | Browser QA review finder, dispatched by /review when a visual change meets a running dev server; the direct /review-browser run is single-pass and does not dispatch it. Preloads the browser criteria a |
| review-code-finder model: opus | Code review finder for /review dispatches and the /review-code direct-run fan-out. Preloads the code criteria and the dispatch contract, reads the project context and file excerpts in the dispatching |
| review-commands-finder model: opus | Slash Command review finder for /review dispatches and the /review-commands direct-run fan-out. Preloads the commands criteria and the dispatch contract, reads the project context and file excerpts in |
| review-copy-finder model: opus | Copy review finder for /review dispatches and the /review-copy direct-run fan-out. Preloads the copy criteria and the dispatch contract, reads the project context and file excerpts in the dispatching |
| review-deps-finder model: opus | Dependency review finder for /review dispatches; the direct /review-deps run is single-pass and does not dispatch it. Preloads the deps criteria and the dispatch contract, reads the project context an |
| review-finder model: inherit | DEPRECATED in 7.0, removed in 8.0 - use the per-kind finders (review-code-finder, review-ux-finder, ...). Generic review specialist worker kept until 8.0 for downstream commands that dispatch it by na |
| review-plan-finder model: opus | Plan Compliance review finder for /review dispatches and the /review-plan direct-run fan-out. Preloads the plan criteria and the dispatch contract, reads the project context and file excerpts in the d |
| review-security-finder model: opus | Security review finder for /review and /review-full dispatches; the direct /review-security run is single-pass and does not dispatch it. Preloads the security criteria and the dispatch contract, reads |
| review-ux-finder model: opus | UX review finder for /review dispatches and the /review-ux direct-run fan-out. Preloads the ux criteria and the dispatch contract, reads the project context and file excerpts in the dispatching prompt |
Slash commands (13)
/ask-gemini/ask-gpt/codebase-to-course/create-issue/create-plan/document/execute/explore/package-review/pair-debug/peer-review/review/worktree
Similar rigs
serionist/omniverse-mcp
MCP server + Kit extension that lets AI assistants drive NVIDIA Omniverse apps (Isaac Sim, USD Composer): inspect and edit USD scenes, capture viewports with segmentation, control robots, physics and recordings.
Fort Knox 9.0k tok · · Q22
serionist/docket
A personal memory and task manager that Claude works in: raw and processed, in your own git repo, local first.
Skill Collector 8.0k tok · · Q19
blackbird987/birdybot
Drive Claude Code CLI from your phone — Discord forum-per-repo, worktree-isolated builds, autonomous build/review/verify/merge chains.
Pragmatist 4.3k tok · · Q5
ThomasTartrau/ironflow
Workflow orchestration engine in Rust. Workflows as imperative code (no YAML/DSL): state machines, parallel execution, approval gates, AI-agent steps. Published on crates.io.
Pragmatist 3.2k tok · · Q0