vamsivarma27/proofloop
Your coding agent says done. ProofLoop makes it prove it. Evidence-gated engineering harness and Claude Code plugin.
field: Developer tools workflow: Planning & specsGit & commitsTesting & TDDResearchSecurity review runs claude in CI
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 vamsivarma27/proofloop/agents ./rig-proofloop/agents $ npx degit vamsivarma27/proofloop/skills ./rig-proofloop/skills $ npx degit vamsivarma27/proofloop/hooks ./rig-proofloop/hooks
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 vamsivarma27/proofloop/skills/approve .claude/skills/approve $ npx degit vamsivarma27/proofloop/skills/bug .claude/skills/bug $ npx degit vamsivarma27/proofloop/skills/cancel .claude/skills/cancel $ npx degit vamsivarma27/proofloop/skills/diagnose .claude/skills/diagnose $ npx degit vamsivarma27/proofloop/skills/feature .claude/skills/feature $ npx degit vamsivarma27/proofloop/skills/ledger .claude/skills/ledger $ npx degit vamsivarma27/proofloop/skills/performance .claude/skills/performance $ npx degit vamsivarma27/proofloop/skills/protocols .claude/skills/protocols $ npx degit vamsivarma27/proofloop/skills/resume .claude/skills/resume $ npx degit vamsivarma27/proofloop/skills/setup .claude/skills/setup $ npx degit vamsivarma27/proofloop/skills/start-project .claude/skills/start-project $ npx degit vamsivarma27/proofloop/skills/status .claude/skills/status $ npx degit vamsivarma27/proofloop/skills/verify .claude/skills/verify
$ curl -fsSL --create-dirs -o .claude/agents/architect.md https://raw.githubusercontent.<redacted>.md $ curl -fsSL --create-dirs -o .claude/agents/browser-evaluator.md https://raw.githubusercontent.com/vamsivarma27/proofloop/main/agents/browser-evaluator.md $ curl -fsSL --create-dirs -o .claude/agents/code-reviewer.md https://raw.githubusercontent.com/vamsivarma27/proofloop/main/agents/code-reviewer.md $ curl -fsSL --create-dirs -o .claude/agents/evidence-evaluator.md https://raw.githubusercontent.com/vamsivarma27/proofloop/main/agents/evidence-evaluator.md $ curl -fsSL --create-dirs -o .claude/agents/implementer.md https://raw.githubusercontent.<redacted>.md $ curl -fsSL --create-dirs -o .claude/agents/product-discovery.md https://raw.githubusercontent.com/vamsivarma27/proofloop/main/agents/product-discovery.md $ curl -fsSL --create-dirs -o .claude/agents/researcher.md https://raw.githubusercontent.<redacted>.md $ curl -fsSL --create-dirs -o .claude/agents/security-reviewer.md https://raw.githubusercontent.com/vamsivarma27/proofloop/main/agents/security-reviewer.md $ curl -fsSL --create-dirs -o .claude/agents/test-designer.md https://raw.githubusercontent.com/vamsivarma27/proofloop/main/agents/test-designer.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 (13)
Subagents (9)
| architect model: inherit | Inspects the real repository and produces impact maps, dependency graphs and test plans, labeling inferred information as inferred. |
| browser-evaluator model: inherit | Plans Playwright journeys with explicit assertions for success, error, empty and permission states. Read-only; reports exact assertions. |
| code-reviewer model: inherit | Fresh-context review of the diff against approved requirements, conventions, API compatibility and data integrity, with file/line evidence. |
| evidence-evaluator model: inherit | Checks that required gates ran on the right commit, artifacts exist, nothing was silently skipped, and the final status matches the evidence. |
| implementer model: inherit | Changes code only inside its assigned worktree and file scope and runs targeted checks. Cannot self-approve verification. |
| product-discovery model: inherit | Finds the real outcome, personas, constraints and measurable success criteria. Asks one focused question at a time; never invents user research. |
| researcher model: inherit | Searches current credible sources and records URL, date consulted, claim supported and uncertainty for every finding. |
| security-reviewer model: inherit | Fresh-context security review: authorization, validation, secrets, injection, dependencies and privilege boundaries. |
| test-designer model: inherit | Writes tests from acceptance criteria and known failure modes before implementation; never weakens assertions to pass. |
Similar rigs
macchi23/penusta-ticket
架空サッカースタジアムのチケット販売サイト(Laravel + Next.js + PostgreSQL の遊び場)
Orchestrator 1.3k tok · · Q0
iadr-dev/greenlight
AI × BDD spec-driven development for Claude Code. greenlight turns a feature request into human-approved, executable specifications, and then into working code. People approve the contract (rules and examples); the AI does the mechanical wo
Orchestrator 1.5k tok · · Q0
DerAlo/ari-auto-pacing
quota-pacer: Claude Code plugin that paces your work so the 5-hour and weekly quota last until the reset
Minimalist 173 tok · · Q0
AdamAkhlaq/claude-code-toolkit
A browsable catalog of Claude Code building blocks — CLAUDE.md, settings, hooks, agents, and doc-writing skills — to copy into your own project and make your own.
Skill Collector 2.5k tok · · Q21