~ / rigs / qyndex / claude-init

qyndex/claude-init

Production-grade Claude Code harness — a spec-driven, TDD-first, verification-gated, security-aware .claude/ scaffold you drop into any repo with one command.

↗ GitHub ★ 2 MIT updated today project Claude Code

share on X

ARCHETYPE

Fort Knox

Deny lists, pre-tool hooks, sandboxing. Nothing touches prod without a signature.

CONTEXT TAX · EVERY TURN

~98.5k tokens

Context hog · median rig: 2.3k · breakdown

GUARDRAILS

5/5

Blocks destructive commands · Protects secrets · Pre-tool screening hook · No YOLO mode · Sandbox or ask-first rules · details

Copy this rig

# review before running: this installs third-party code
$ claude mcp add filesystem -- npx -y @modelcontextprotocol/server-filesystem@2026.1.14 .
$ claude mcp add git -- uvx mcp-server-git==2026.1.14 --repository .
$ claude mcp add --transport http github https://api.githubcopilot.com/mcp/ -H 'Authorization: YOUR_VALUE'
$ claude mcp add chrome-devtools -- npx -y chrome-devtools-mcp@0.21.0
$ claude mcp add playwright -e HEADLESS=YOUR_HEADLESS -- npx -y @playwright/mcp@1.50.0
$ claude mcp add context7 -- npx -y @upstash/context7-mcp@1.4.0
$ claude mcp add postgres -e DATABASE_URI=YOUR_DATABASE_URI -- uvx postgres-mcp --access-mode=restricted
$ claude mcp add --transport http sentry https://mcp.sentry.dev/mcp
$ claude mcp add --transport http apify https://mcp.apify.com
$ claude mcp add semgrep -- semgrep mcp
$ claude mcp add scheduled-tasks -- npx -y @anthropic-ai/mcp-scheduled-tasks@0.4.0
$ claude mcp add openfeature -e OPENFEATURE_PROVIDER=YOUR_OPENFEATURE_PROVIDER -- npx -y @openfeature/mcp-server
$ claude mcp add --transport http posthog https://mcp.posthog.com/sse -H 'Authorization: YOUR_VALUE'
$ claude mcp add graphify -- uvx --from 'graphify[mcp]==0.8.21' python -m graphify.serve
$ claude mcp add langfuse -e LANGFUSE_BASEURL=YOUR_LANGFUSE_BASEURL -e LANGFUSE_PUBLIC_KEY=YOUR_LANGFUSE_PUBLIC_KEY -e LANGFUSE_SECRET_KEY=YOUR_LANGFUSE_SECRET_KEY -- npx -y @langfuse/mcp-server-langfuse@0.4.0
$ claude mcp add fireflies -e FIREFLIES_<redacted> -- npx -y @fireflies/mcp-server@0.2.0
$ claude mcp add intercom -e INTERCOM_ACCESS_<redacted> -- npx -y @intercom/mcp-server@0.3.0
$ claude mcp add pendo -e PENDO_INTEGRATION_KEY=YOUR_PENDO_INTEGRATION_KEY -- npx -y @pendo/mcp-server@0.1.0
$ claude mcp add --transport http replicate https://mcp.replicate.com/
$ claude mcp add recraft -e RECRAFT_<redacted> -- npx -y @recraft-ai/mcp-server@0.1.0
$ claude mcp add pexels -e PEXELS_<redacted> -- npx -y pexels-mcp-server@0.2.0
$ claude mcp add osv -- uvx --from git+https://github.com/StacklokLabs/osv-mcp@v0.1.0 osv-mcp
$ npx degit qyndex/claude-init/.claude ./rig-claude-init  # 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 qyndex/claude-init/.claude/skills/analyze .claude/skills/analyze
$ npx degit qyndex/claude-init/.claude/skills/animation-discipline .claude/skills/animation-discipline
$ npx degit qyndex/claude-init/.claude/skills/api-versioning .claude/skills/api-versioning
$ npx degit qyndex/claude-init/.claude/skills/autopilot .claude/skills/autopilot
$ npx degit qyndex/claude-init/.claude/skills/browser-e2e .claude/skills/browser-e2e
$ npx degit qyndex/claude-init/.claude/skills/characterize .claude/skills/characterize
$ npx degit qyndex/claude-init/.claude/skills/clarify .claude/skills/clarify
$ npx degit qyndex/claude-init/.claude/skills/codemod .claude/skills/codemod
$ npx degit qyndex/claude-init/.claude/skills/constitution-compact .claude/skills/constitution-compact
$ npx degit qyndex/claude-init/.claude/skills/constitution .claude/skills/constitution
$ npx degit qyndex/claude-init/.claude/skills/context-budget .claude/skills/context-budget
$ npx degit qyndex/claude-init/.claude/skills/creative-web .claude/skills/creative-web
$ npx degit qyndex/claude-init/.claude/skills/db-migration .claude/skills/db-migration
$ npx degit qyndex/claude-init/.claude/skills/debug .claude/skills/debug
$ npx degit qyndex/claude-init/.claude/skills/detect-stack .claude/skills/detect-stack
$ npx degit qyndex/claude-init/.claude/skills/dispatch-criteria .claude/skills/dispatch-criteria
$ npx degit qyndex/claude-init/.claude/skills/dream .claude/skills/dream
$ npx degit qyndex/claude-init/.claude/skills/experiment .claude/skills/experiment
$ npx degit qyndex/claude-init/.claude/skills/fastapi .claude/skills/fastapi
$ npx degit qyndex/claude-init/.claude/skills/flag-cleanup .claude/skills/flag-cleanup
$ npx degit qyndex/claude-init/.claude/skills/flag-rollout .claude/skills/flag-rollout
$ npx degit qyndex/claude-init/.claude/skills/flask-realtime .claude/skills/flask-realtime
$ npx degit qyndex/claude-init/.claude/skills/framework-detector .claude/skills/framework-detector
$ npx degit qyndex/claude-init/.claude/skills/front-end-design .claude/skills/front-end-design
$ npx degit qyndex/claude-init/.claude/skills/grill-me .claude/skills/grill-me
$ npx degit qyndex/claude-init/.claude/skills/handoff .claude/skills/handoff
$ npx degit qyndex/claude-init/.claude/skills/image-generation .claude/skills/image-generation
$ npx degit qyndex/claude-init/.claude/skills/implement .claude/skills/implement
$ npx degit qyndex/claude-init/.claude/skills/initiative .claude/skills/initiative
$ npx degit qyndex/claude-init/.claude/skills/instinct .claude/skills/instinct
$ npx degit qyndex/claude-init/.claude/skills/loop .claude/skills/loop
$ npx degit qyndex/claude-init/.claude/skills/nextjs .claude/skills/nextjs
$ npx degit qyndex/claude-init/.claude/skills/observability .claude/skills/observability
$ npx degit qyndex/claude-init/.claude/skills/okr-align .claude/skills/okr-align
$ npx degit qyndex/claude-init/.claude/skills/parallel-swarm .claude/skills/parallel-swarm
$ npx degit qyndex/claude-init/.claude/skills/pivot .claude/skills/pivot
$ npx degit qyndex/claude-init/.claude/skills/plan .claude/skills/plan
$ npx degit qyndex/claude-init/.claude/skills/pr-body .claude/skills/pr-body
$ npx degit qyndex/claude-init/.claude/skills/prime-discipline .claude/skills/prime-discipline
$ npx degit qyndex/claude-init/.claude/skills/prod-observability .claude/skills/prod-observability

$ curl -fsSL --create-dirs -o .claude/agents/architect.md https://raw.githubusercontent.com/qyndex/claude-init/main/.claude/agents/core/architect.md
$ curl -fsSL --create-dirs -o .claude/agents/coordinator.md https://raw.githubusercontent.com/qyndex/claude-init/main/.claude/agents/core/coordinator.md
$ curl -fsSL --create-dirs -o .claude/agents/feature-stream.md https://raw.githubusercontent.com/qyndex/claude-init/main/.claude/agents/core/feature-stream.md
$ curl -fsSL --create-dirs -o .claude/agents/implementer.md https://raw.githubusercontent.com/qyndex/claude-init/main/.claude/agents/core/implementer.md
$ curl -fsSL --create-dirs -o .claude/agents/planner.md https://raw.githubusercontent.com/qyndex/claude-init/main/.claude/agents/core/planner.md
$ curl -fsSL --create-dirs -o .claude/agents/roadmap-architect.md https://raw.githubusercontent.com/qyndex/claude-init/main/.claude/agents/core/roadmap-architect.md
$ curl -fsSL --create-dirs -o .claude/agents/anti-slop-reviewer.md https://raw.githubusercontent.com/qyndex/claude-init/main/.claude/agents/quality/anti-slop-reviewer.md
$ curl -fsSL --create-dirs -o .claude/agents/reviewer.md https://raw.githubusercontent.com/qyndex/claude-init/main/.claude/agents/quality/reviewer.md
$ curl -fsSL --create-dirs -o .claude/agents/security.md https://raw.githubusercontent.com/qyndex/claude-init/main/.claude/agents/quality/security.md
$ curl -fsSL --create-dirs -o .claude/agents/tester.md https://raw.githubusercontent.com/qyndex/claude-init/main/.claude/agents/quality/tester.md
$ curl -fsSL --create-dirs -o .claude/agents/verifier.md https://raw.githubusercontent.com/qyndex/claude-init/main/.claude/agents/quality/verifier.md
$ curl -fsSL --create-dirs -o .claude/agents/debugger.md https://raw.githubusercontent.com/qyndex/claude-init/main/.claude/agents/specialists/debugger.md
$ curl -fsSL --create-dirs -o .claude/agents/designer.md https://raw.githubusercontent.com/qyndex/claude-init/main/.claude/agents/specialists/designer.md
$ curl -fsSL --create-dirs -o .claude/agents/doc-writer.md https://raw.githubusercontent.com/qyndex/claude-init/main/.claude/agents/specialists/doc-writer.md
$ curl -fsSL --create-dirs -o .claude/agents/extractor.md https://raw.githubusercontent.com/qyndex/claude-init/main/.claude/agents/specialists/extractor.md
$ curl -fsSL --create-dirs -o .claude/agents/feedback-extractor.md https://raw.githubusercontent.com/qyndex/claude-init/main/.claude/agents/specialists/feedback-extractor.md
$ curl -fsSL --create-dirs -o .claude/agents/release.md https://raw.githubusercontent.com/qyndex/claude-init/main/.claude/agents/specialists/release.md
$ curl -fsSL --create-dirs -o .claude/agents/researcher.md https://raw.githubusercontent.com/qyndex/claude-init/main/.claude/agents/specialists/researcher.md
$ curl -fsSL --create-dirs -o .claude/agents/architect.md https://raw.githubusercontent.com/qyndex/claude-init/main/verify/2026-06-12-e2e-fixes/staged/.claude/agents/core/architect.md
$ curl -fsSL --create-dirs -o .claude/agents/coordinator.md https://raw.githubusercontent.com/qyndex/claude-init/main/verify/2026-06-12-e2e-fixes/staged/.claude/agents/core/coordinator.md
$ curl -fsSL --create-dirs -o .claude/agents/feature-stream.md https://raw.githubusercontent.com/qyndex/claude-init/main/verify/2026-06-12-e2e-fixes/staged/.claude/agents/core/feature-stream.md
$ curl -fsSL --create-dirs -o .claude/agents/implementer.md https://raw.githubusercontent.com/qyndex/claude-init/main/verify/2026-06-12-e2e-fixes/staged/.claude/agents/core/implementer.md
$ curl -fsSL --create-dirs -o .claude/agents/planner.md https://raw.githubusercontent.com/qyndex/claude-init/main/verify/2026-06-12-e2e-fixes/staged/.claude/agents/core/planner.md
$ curl -fsSL --create-dirs -o .claude/agents/verifier.md https://raw.githubusercontent.com/qyndex/claude-init/main/verify/2026-06-12-e2e-fixes/staged/.claude/agents/quality/verifier.md
$ curl -fsSL --create-dirs -o .claude/agents/feedback-extractor.md https://raw.githubusercontent.com/qyndex/claude-init/main/verify/2026-06-12-e2e-fixes/staged/.claude/agents/specialists/feedback-extractor.md
$ curl -fsSL --create-dirs -o .claude/agents/release.md https://raw.githubusercontent.com/qyndex/claude-init/main/verify/2026-06-12-e2e-fixes/staged/.claude/agents/specialists/release.md
$ curl -fsSL --create-dirs -o .claude/agents/researcher.md https://raw.githubusercontent.com/qyndex/claude-init/main/verify/2026-06-12-e2e-fixes/staged/.claude/agents/specialists/researcher.md
$ curl -fsSL --create-dirs -o .claude/agents/architect.md https://raw.githubusercontent.com/qyndex/claude-init/main/verify/2026-06-12-gap-fixes/staged/.claude/agents/core/architect.md
$ curl -fsSL --create-dirs -o .claude/agents/implementer.md https://raw.githubusercontent.com/qyndex/claude-init/main/verify/2026-06-12-gap-fixes/staged/.claude/agents/core/implementer.md
$ curl -fsSL --create-dirs -o .claude/agents/roadmap-architect.md https://raw.githubusercontent.com/qyndex/claude-init/main/verify/2026-06-12-gap-fixes/staged/.claude/agents/core/roadmap-architect.md
$ curl -fsSL --create-dirs -o .claude/agents/reviewer.md https://raw.githubusercontent.com/qyndex/claude-init/main/verify/2026-06-12-gap-fixes/staged/.claude/agents/quality/reviewer.md
$ curl -fsSL --create-dirs -o .claude/agents/security.md https://raw.githubusercontent.com/qyndex/claude-init/main/verify/2026-06-12-gap-fixes/staged/.claude/agents/quality/security.md
$ curl -fsSL --create-dirs -o .claude/agents/debugger.md https://raw.githubusercontent.com/qyndex/claude-init/main/verify/2026-06-12-gap-fixes/staged/.claude/agents/specialists/debugger.md

Merge into .claude/settings.json (project) or ~/.claude/settings.json (user). Hook commands reference scripts in the source repo — copy those too.

{
  "permissions": {
    "deny": [
      "Bash(DROP DATABASE:*)",
      "Bash(DROP TABLE:*)",
      "Bash(TRUNCATE TABLE:*)",
      "Bash(dd if=:*)",
      "Bash(git branch -D main)",
      "Bash(git branch -D master)",
      "Bash(git checkout -- .)",
      "Bash(git clean -fdx)",
      "Bash(git commit --no-verify:*)",
      "Bash(git push --force origin main)",
      "Bash(git push --force origin master)",
      "Bash(git push --no-verify:*)",
      "Bash(git push -f origin main)",
      "Bash(git push -f origin master)",
      "Bash(git reset --hard origin/main)",
      "Bash(git reset --hard origin/master)",
      "Bash(mkfs:*)",
      "Bash(rm -rf $HOME)",
      "Bash(rm -rf /)",
      "Bash(rm -rf /*)",
      "Bash(rm -rf ~)",
      "Bash(rm -rf ~/*)",
      "Bash(sudo rm:*)",
      "Read(**/*.kdbx)",
      "Read(**/*.key)",
      "Read(**/*.kubeconfig)",
      "Read(**/*.p12)",
      "Read(**/*.pem)",
      "Read(**/*.pfx)",
      "Read(**/*.tfstate)",
      "Read(**/*.tfvars)",
      "Read(**/*_rsa)",
      "Read(**/*credentials*)",
      "Read(**/.aws/credentials)",
      "Read(**/.docker/config.json)",
      "Read(**/.env)",
      "Read(**/.env.*)",
      "Read(**/.git-credentials)",
      "Read(**/.gnupg/**)",
      "Read(**/.netrc)",
      "Read(**/.npmrc)",
      "Read(**/.pypirc)",
      "Read(**/.ssh/**)",
      "Read(**/_netrc)",
      "Read(**/id_rsa*)",
      "Read(**/known_hosts)",
      "Read(**/kubeconfig)",
      "Read(**/secrets/**)",
      "Read(./.env)",
      "Read(./.env.*)",
      "Write(**/*.kdbx)",
      "Write(**/*.key)",
      "Write(**/*.kubeconfig)",
      "Write(**/*.p12)",
      "Write(**/*.pem)",
      "Write(**/*.pfx)",
      "Write(**/*.tfstate)",
      "Write(**/*.tfvars)",
      "Write(**/*_rsa)",
      "Write(**/*credentials*)",
      "Write(**/.docker/config.json)",
      "Write(**/.env)",
      "Write(**/.env.*)",
      "Write(**/.git-credentials)",
      "Write(**/.gnupg/**)",
      "Write(**/.netrc)",
      "Write(**/.npmrc)",
      "Write(**/.pypirc)",
      "Write(**/.ssh/**)",
      "Write(**/_netrc)",
      "Write(**/id_rsa*)",
      "Write(**/kubeconfig)",
      "Write(./.env)",
      "Write(./.env.*)",
      "Bash(DROP DATABASE:*)",
      "Bash(DROP TABLE:*)",
      "Bash(TRUNCATE TABLE:*)",
      "Bash(dd if=:*)",
      "Bash(git branch -D main)",
      "Bash(git branch -D master)"
    ],
    "ask": [
      "Bash(git push:*)",
      "Bash(git push --force-with-lease:*)",
      "Bash(gh pr merge:*)",
      "Bash(gh release create:*)",
      "Bash(npm publish:*)",
      "Bash(pnpm publish:*)",
      "Bash(cargo publish:*)",
      "Bash(docker push:*)",
      "Bash(terraform apply:*)",
      "Bash(terraform destroy:*)",
      "Bash(kubectl apply:*)",
      "Bash(kubectl delete:*)",
      "Bash(aws:*)",
      "Bash(vercel deploy:*)",
      "Bash(flyctl deploy:*)",
      "Bash(railway deploy:*)",
      "Bash(rm -rf:*)",
      "Bash(claude -p:*)",
      "Bash(claude --bg:*)",
      "Bash(claude --remote:*)",
      "Bash(git push:*)",
      "Bash(git push --force-with-lease:*)",
      "Bash(gh pr merge:*)",
      "Bash(gh release create:*)",
      "Bash(npm publish:*)",
      "Bash(pnpm publish:*)",
      "Bash(cargo publish:*)",
      "Bash(docker push:*)",
      "Bash(terraform apply:*)",
      "Bash(terraform destroy:*)",
      "Bash(kubectl apply:*)",
      "Bash(kubectl delete:*)",
      "Bash(aws:*)",
      "Bash(vercel deploy:*)",
      "Bash(flyctl deploy:*)",
      "Bash(railway deploy:*)",
      "Bash(rm -rf:*)",
      "Bash(claude -p:*)",
      "Bash(claude --bg:*)",
      "Bash(claude --remote:*)"
    ]
  },
  "hooks": {
    "PreToolUse": [
      {
        "matcher": "Bash",
        "hooks": [
          {
            "type": "command",
            "command": "bash .claude/hooks/pre-bash-guard.sh"
          },
          {
            "type": "command",
            "command": "bash .claude/hooks/pre-spawn-cost-gate.sh"
          },
          {
            "type": "command",
            "command": "bash .claude/hooks/pre-bash-dep-freshness.sh"
          },
          {
            "type": "command",
            "command": "bash .claude/hooks/pre-bash-guard.sh"
          },
          {
            "type": "command",
            "command": "bash .claude/hooks/pre-spawn-cost-gate.sh"
          },
          {
            "type": "command",
            "command": "bash .claude/hooks/pre-bash-dep-freshness.sh"
          }
        ]
      },
      {
        "matcher": "Write|Edit|NotebookEdit",
        "hooks": [
          {
            "type": "command",
            "command": "bash .claude/hooks/pre-edit-constitution-guard.sh"
          },
          {
            "type": "command",
            "command": "bash .claude/hooks/pre-write-secret-scan.sh"
          },
          {
            "type": "command",
            "command": "bash .claude/hooks/pre-edit-legacy-guard.sh"
          },
          {
            "type": "command",
            "command": "bash .claude/hooks/pre-lane-guard.sh"
          },
          {
            "type": "command",
            "command": "bash .claude/hooks/pre-edit-constitution-guard.sh"
          },
          {
            "type": "command",
            "command": "bash .claude/hooks/pre-write-secret-scan.sh"
          },
          {
            "type": "command",
            "command": "bash .claude/hooks/pre-edit-legacy-guard.sh"
          },
          {
            "type": "command",
            "command": "bash .claude/hooks/pre-lane-guard.sh"
          }
        ]
      },
      {
        "matcher": "Agent|Task",
        "hooks": [
          {
            "type": "command",
            "command": "bash .claude/hooks/pre-spawn-cost-gate.sh"
          },
          {
            "type": "command",
            "command": "bash .claude/hooks/subagent-context.sh"
          },
          {
            "type": "command",
            "command": "bash .claude/hooks/pre-spawn-cost-gate.sh"
          },
          {
            "type": "command",
            "command": "bash .claude/hooks/subagent-context.sh"
          }
        ]
      },
      {
        "matcher": "mcp__filesystem__write_file|mcp__filesystem__edit_file|mcp__filesystem__move_file|mcp__github__create_or_update_file|mcp",
        "hooks": [
          {
            "type": "command",
            "command": "bash .claude/hooks/pre-mcp-write-guard.sh"
          },
          {
            "type": "command",
            "command": "bash .claude/hooks/pre-mcp-write-guard.sh"
          }
        ]
      }
    ]
  }
}

MCP servers (22)

server	source	est. tokens
Filesystem	npm	4.2k
Git	pypi	3.2k
GitHub MCP · "github"	remote · remote	18.0k
Chrome DevTools MCP · "chrome-devtools"	npm	9.0k
Playwright MCP · "playwright" env: HEADLESS	npm	7.5k
Context7	npm	1.2k
postgres-mcp · "postgres" env: DATABASE_URI	pypi	2.5k
Sentry	remote · remote	4.5k
mcp.apify.com · "apify"	remote · remote	2.5k
semgrep	binary	2.5k
@anthropic-ai/mcp-scheduled-tasks · "scheduled-tasks"	npm	2.5k
@openfeature/mcp-server · "openfeature" env: OPENFEATURE_PROVIDER	npm	2.5k
mcp.posthog.com · "posthog"	remote · remote	2.5k
graphify[mcp] · "graphify"	pypi	2.5k
@langfuse/mcp-server-langfuse · "langfuse" env: LANGFUSE_BASEURL, LANGFUSE_PUBLIC_KEY, LANGFUSE_SECRET_KEY	npm	2.5k
@fireflies/mcp-server · "fireflies" env: FIREFLIES_API_KEY	npm	2.5k
@intercom/mcp-server · "intercom" env: INTERCOM_ACCESS_TOKEN	npm	2.5k
@pendo/mcp-server · "pendo" env: PENDO_INTEGRATION_KEY	npm	2.5k
mcp.replicate.com · "replicate"	remote · remote	2.5k
@recraft-ai/mcp-server · "recraft" env: RECRAFT_API_KEY	npm	2.5k
pexels-mcp-server · "pexels" env: PEXELS_API_KEY	npm	2.5k
stackloklabs/osv-mcp · "osv"	git	2.5k

Skills (58)

Subagents (33)

architect model: opus	Use proactively at the start of any non-trivial feature, refactor, or system change. Designs the system, authors specs and plans, picks technologies, defines acceptance criteria. Read-only — never wri
coordinator model: sonnet	Orchestrates a swarm of feature-stream background sessions. Plans streams from tasks/TASKS.md, dispatches each as `claude --bg --worktree feat-<N>`, monitors via `claude agents --json`, merges PRs as
feature-stream model: sonnet	The agent running inside each `claude --bg --worktree feat-<N>` session. Reads its brief, executes the assigned tasks via strict TDD, in-stream subagents (implementer + spec-reviewer + code-quality-re
implementer model: opus	Use to execute a single task from tasks/TASKS.md. Writes code following the approved plan, using strict TDD (red→green→refactor). Reads the spec, plan, and task; writes tests first; implements minimal
planner model: sonnet	Use after architect produces a plan.md. Decomposes a plan into 2-5 minute machine-verifiable tasks with explicit dependencies, parallel-safe markers [P], and acceptance commands. Read-only except for
roadmap-architect model: sonnet	Use for multi-quarter program design — initiative authoring, OKR alignment, roadmap cut/promote, phase decomposition that bridges specs to objectives. Different from the `architect` agent (per-feature
anti-slop-reviewer model: sonnet	Use to triage an incoming PR or issue into exactly one of eight classes — actionable-bug, actionable-docs, actionable-feature, duplicate, spam, generated-slop, security-sensitive, not-reproducible. Re
reviewer model: opus	Use after implementation but before /ship. Reviews the diff for correctness, readability, maintainability, performance, and adherence to project conventions. Read-only — posts comments only. Loops wit
security model: opus	Use proactively before /ship on any change touching auth, data handling, network, secrets, dependencies, or user input. Performs a deep security review using the Anthropic claude-code-security-review
tester model: sonnet	Use to expand test coverage, add edge-case tests, write integration or E2E tests, or run mutation testing against existing code. Different from implementer's TDD loop — the tester audits and extends,
verifier model: sonnet	Use after implementation and before /ship. Runs the actual application end-to-end, exercises the user journey from the spec, captures evidence (screenshots, logs, HTTP traces), and confirms the change
debugger model: opus	Use when a test fails unexpectedly, the app crashes, behavior diverges from expected, or any time the cause is non-obvious. Runs the Superpowers systematic-debugging four-phase loop: reproduce → isola
designer model: opus	Use when a spec calls for visual craft — landing pages, hero sections, redesigns, "make it look amazing". Owns front-end-design + tailwind-discipline + typography + animation-discipline + image-genera
doc-writer model: sonnet	Use after a feature ships, after an ADR is decided, or when README/CHANGELOG/runbook drifts from reality. Updates docs to match what the code actually does. Never invents behavior — reads the code fir
extractor model: opus	Use when an initiative is being abandoned (/abandon) — extracts retained learnings before the files move to archive. Reads child specs + plans + ADRs + git log; writes post-mortem + anti-patterns + sa
feedback-extractor model: sonnet	Use to pull customer signal out of sales-call transcripts, support tickets, NPS surveys, Pendo polls, and product analytics. Read-only on external sources AND the repo — returns structured feedback en
release model: sonnet	Use as the final agent in /ship. Confirms all quality gates pass, opens the PR, watches CI, merges (squash) once green, tags the release, and triggers deploy. Stops at any gate failure and reports.
researcher model: sonnet	Use when a decision needs evidence the local codebase can't provide — picking between libraries, evaluating an architectural pattern, comparing APIs, surveying competitor implementations, reading offi
architect model: opus	Use proactively at the start of any non-trivial feature, refactor, or system change. Designs the system, authors specs and plans, picks technologies, defines acceptance criteria. Read-only — never wri
coordinator model: sonnet	Orchestrates a swarm of feature-stream background sessions. Plans streams from tasks/TASKS.md, dispatches each as `claude --bg --worktree feat-<N>`, monitors via `claude agents --json`, merges PRs as
feature-stream model: sonnet	The agent running inside each `claude --bg --worktree feat-<N>` session. Reads its brief, executes the assigned tasks via strict TDD, in-stream subagents (implementer + spec-reviewer + code-quality-re
implementer model: opus	Use to execute a single task from tasks/TASKS.md. Writes code following the approved plan, using strict TDD (red→green→refactor). Reads the spec, plan, and task; writes tests first; implements minimal
planner model: sonnet	Use after architect produces a plan.md. Decomposes a plan into 2-5 minute machine-verifiable tasks with explicit dependencies, parallel-safe markers [P], and acceptance commands. Read-only except for
verifier model: sonnet	Use after implementation and before /ship. Runs the actual application end-to-end, exercises the user journey from the spec, captures evidence (screenshots, logs, HTTP traces), and confirms the change
feedback-extractor model: sonnet	Use to pull customer signal out of sales-call transcripts, support tickets, NPS surveys, Pendo polls, and product analytics. Read-only on external sources AND the repo — returns structured feedback en
release model: sonnet	Use as the final agent in /ship. Confirms all quality gates pass, opens the PR, watches CI, merges (squash) once green, tags the release, and triggers deploy. Stops at any gate failure and reports.
researcher model: sonnet	Use when a decision needs evidence the local codebase can't provide — picking between libraries, evaluating an architectural pattern, comparing APIs, surveying competitor implementations, reading offi
architect model: opus	Use proactively at the start of any non-trivial feature, refactor, or system change. Designs the system, authors specs and plans, picks technologies, defines acceptance criteria. Read-only — never wri
implementer model: opus	Use to execute a single task from tasks/TASKS.md. Writes code following the approved plan, using strict TDD (red→green→refactor). Reads the spec, plan, and task; writes tests first; implements minimal
roadmap-architect model: sonnet	Use for multi-quarter program design — initiative authoring, OKR alignment, roadmap cut/promote, phase decomposition that bridges specs to objectives. Different from the `architect` agent (per-feature
reviewer model: opus	Use after implementation but before /ship. Reviews the diff for correctness, readability, maintainability, performance, and adherence to project conventions. Read-only — posts comments only. Loops wit
security model: opus	Use proactively before /ship on any change touching auth, data handling, network, secrets, dependencies, or user input. Performs a deep security review using the Anthropic claude-code-security-review
debugger model: opus	Use when a test fails unexpectedly, the app crashes, behavior diverges from expected, or any time the cause is non-obvious. Runs the Superpowers systematic-debugging four-phase loop: reproduce → isola

Hooks (60)

event	matcher	runs
SessionStart	startup\|resume\|clear\|compact	bash .claude/hooks/session-start.sh
SessionStart	startup\|resume\|clear\|compact	bash .claude/hooks/session-start-context.sh
SessionStart	startup\|resume\|clear\|compact	bash .claude/scripts/check-overnight-fired.sh
UserPromptSubmit	*	bash .claude/hooks/user-prompt-context.sh
UserPromptSubmit	*	bash .claude/hooks/workflow-state.sh
UserPromptSubmit	*	bash .claude/hooks/skill-router.sh
UserPromptSubmit	*	bash .claude/hooks/session-heartbeat.sh
UserPromptSubmit	*	bash .claude/hooks/context-monitor.sh
PreToolUse	Bash	bash .claude/hooks/pre-bash-guard.sh
PreToolUse	Bash	bash .claude/hooks/pre-spawn-cost-gate.sh
PreToolUse	Bash	bash .claude/hooks/pre-bash-dep-freshness.sh
PreToolUse	Write\|Edit\|NotebookEdit	bash .claude/hooks/pre-edit-constitution-guard.sh
PreToolUse	Write\|Edit\|NotebookEdit	bash .claude/hooks/pre-write-secret-scan.sh
PreToolUse	Write\|Edit\|NotebookEdit	bash .claude/hooks/pre-edit-legacy-guard.sh
PreToolUse	Write\|Edit\|NotebookEdit	bash .claude/hooks/pre-lane-guard.sh
PreToolUse	Agent\|Task	bash .claude/hooks/pre-spawn-cost-gate.sh
PreToolUse	Agent\|Task	bash .claude/hooks/subagent-context.sh
PreToolUse	mcp__filesystem__write_file\|mcp__filesystem__edit_file\|mcp__filesystem__move_file\|mcp__github__create_or_update_file\|mcp	bash .claude/hooks/pre-mcp-write-guard.sh
PostToolUse	Write\|Edit	bash .claude/hooks/post-write-format.sh
PostToolUse	Write\|Edit	bash .claude/hooks/instinct-observer.sh
PostToolUse	Write\|Edit	bash .claude/hooks/post-write-roadmap.sh
PostToolUse	Bash	bash .claude/hooks/post-bash-log.sh
PostToolUse	Bash	bash .claude/hooks/instinct-observer.sh
PostToolUse	Skill	bash .claude/hooks/skill-use-log.sh
PostToolUse	Read	bash .claude/hooks/read-volume-nudge.sh
Stop	*	bash .claude/hooks/stop-verify.sh
Stop	*	bash .claude/hooks/auto-dream-check.sh
Stop	*	bash .claude/hooks/ta<redacted>.sh
SubagentStop	*	bash .claude/hooks/subagent-stop.sh
PreCompact	auto\|manual	bash .claude/hooks/pre-compact-witness.sh

Slash commands (55)

/abandon/adopt/adr-walk/audit-plugins/audit-trail/bisect-perf/cleanup/codify-rule/context-pack/create-agent/create-skill/debt/deprecate-api/deps-audit/diagram/dr-drill/dream-review/feedback/flag/handoff/harness-doctor/health/incident-end/incident-start/initiative/instinct/issues/kickoff/lesson-learned/ml-eval/okrs/onboard/oncall-handoff/owner-walk/perf-trend/pii-audit/pivot/prime/roadmap/rollback-flag/seed-patterns/spec-drift-check/spec/status/sunset/supersede/swarm/abort/swarm/dispatch/swarm/merge/swarm/plan/swarm/respawn/swarm/status/swarm/stop/triage/upgrade

Permissions

deny (120)

Bash(DROP DATABASE:*)

Bash(DROP TABLE:*)

Bash(TRUNCATE TABLE:*)

Bash(dd if=:*)

Bash(git branch -D main)

Bash(git branch -D master)

Bash(git checkout -- .)

Bash(git clean -fdx)

Bash(git commit --no-verify:*)

Bash(git push --force origin main)

Bash(git push --force origin master)

Bash(git push --no-verify:*)

Bash(git push -f origin main)

Bash(git push -f origin master)

Bash(git reset --hard origin/main)

Bash(git reset --hard origin/master)

Bash(mkfs:*)

Bash(rm -rf $HOME)

Bash(rm -rf /)

Bash(rm -rf /*)

Bash(rm -rf ~)

Bash(rm -rf ~/*)

Bash(sudo rm:*)

Read(**/*.kdbx)

Read(**/*.key)

Read(**/*.kubeconfig)

Read(**/*.p12)

Read(**/*.pem)

Read(**/*.pfx)

Read(**/*.tfstate)

Read(**/*.tfvars)

Read(**/*_rsa)

Read(**/*credentials*)

Read(**/.aws/credentials)

Read(**/.docker/config.json)

Read(**/.env)

Read(**/.env.*)

Read(**/.git-credentials)

Read(**/.gnupg/**)

Read(**/.netrc)

Read(**/.npmrc)

Read(**/.pypirc)

Read(**/.ssh/**)

Read(**/_netrc)

Read(**/id_rsa*)

Read(**/known_hosts)

Read(**/kubeconfig)

Read(**/secrets/**)

Read(./.env)

Read(./.env.*)

Write(**/*.kdbx)

Write(**/*.key)

Write(**/*.kubeconfig)

Write(**/*.p12)

Write(**/*.pem)

Write(**/*.pfx)

Write(**/*.tfstate)

Write(**/*.tfvars)

Write(**/*_rsa)

Write(**/*credentials*)

ask (40)

Bash(git push:*)

Bash(git push --force-with-lease:*)

Bash(gh pr merge:*)

Bash(gh release create:*)

Bash(npm publish:*)

Bash(pnpm publish:*)

Bash(cargo publish:*)

Bash(docker push:*)

Bash(terraform apply:*)

Bash(terraform destroy:*)

Bash(kubectl apply:*)

Bash(kubectl delete:*)

Bash(aws:*)

Bash(vercel deploy:*)

Bash(flyctl deploy:*)

Bash(railway deploy:*)

Bash(rm -rf:*)

Bash(claude -p:*)

Bash(claude --bg:*)

Bash(claude --remote:*)

Bash(git push:*)

Bash(git push --force-with-lease:*)

Bash(gh pr merge:*)

Bash(gh release create:*)

Bash(npm publish:*)

Bash(pnpm publish:*)

Bash(cargo publish:*)

Bash(docker push:*)

Bash(terraform apply:*)

Bash(terraform destroy:*)

Bash(kubectl apply:*)

Bash(kubectl delete:*)

Bash(aws:*)

Bash(vercel deploy:*)

Bash(flyctl deploy:*)

Bash(railway deploy:*)

Bash(rm -rf:*)

Bash(claude -p:*)

Bash(claude --bg:*)

Bash(claude --remote:*)

allow (120)

Edit(src/**)

Edit(tests/**)

Edit(specs/**)

Edit(plans/**)

Edit(tasks/**)

Edit(docs/**)

Edit(verify/**)

Edit(.swarms/**)

Edit(initiatives/**)

Edit(.claude/memory/**)

Edit(.claude/memory.proposed/**)

Edit(.claude/state/**)

Edit(.claude/rules/**)

Edit(OVERNIGHT_REPORT.md)

Edit(ADOPTION-REPORT.md)

Edit(roadmap.md)

Edit(OKRs.md)

Read

Glob

Grep

Write(src/**)

Write(tests/**)

Write(specs/**)

Write(plans/**)

Write(tasks/**)

Write(docs/**)

Write(verify/**)

Write(.swarms/**)

Write(.claude/memory/**)

Write(.claude/memory.proposed/**)

Write(.claude/state/**)

Write(.claude/hooks/.log/**)

TodoWrite

WebSearch

WebFetch(domain:docs.anthropic.com)

WebFetch(domain:code.claude.com)

WebFetch(domain:claude.com)

WebFetch(domain:platform.claude.com)

WebFetch(domain:github.com)

WebFetch(domain:raw.githubusercontent.com)

WebFetch(domain:gist.github.com)

WebFetch(domain:developer.mozilla.org)

WebFetch(domain:nodejs.org)

WebFetch(domain:python.org)

WebFetch(domain:docs.python.org)

WebFetch(domain:pypi.org)

WebFetch(domain:npmjs.com)

WebFetch(domain:agentskills.io)

WebFetch(domain:modelcontextprotocol.io)

WebFetch(domain:registry.modelcontextprotocol.io)

WebFetch(domain:apify.com)

WebFetch(domain:docs.apify.com)

Bash(ls:*)

Bash(pwd)

Bash(echo:*)

Bash(date:*)

Bash(which:*)

Bash(whoami)

Bash(uname:*)

Bash(file:*)

Context tax

Instructions: 6.8kMCP tool schemas (est.): 85.1kSkill metadata: 4.0kSubagent metadata: 2.6k

ALWAYS-LOADED FILES (claude-code)

≡

CLAUDE.md · 236 lines · ~3.5k tok

≡

.claude/CLAUDE.md · 213 lines · ~3.3k tok

Estimate: characters ÷ 4 for instruction files and skill/subagent descriptions; MCP schema sizes from typical tool counts. Harnesses with deferred tool loading pay less. Method

Guardrails 5/5

✓

Blocks destructive commands

52 deny/ask rule(s) e.g. Bash(DROP DATABASE:*)

✓

Protects secrets

Denies reads like Read(**/*.key)

✓

Pre-tool screening hook

20 PreToolUse hook(s)

✓

No YOLO mode

Permission prompts stay on

✓

Sandbox or ask-first rules

40 ask rule(s), e.g. Bash(git push:*)

Badge

Add to your README:

[![AgentRigs](https://agentrigs.dev/badge/rig/qyndex/claude-init.svg)](https://agentrigs.dev/r/qyndex/claude-init)

Indexed from commit 148238ad8950 on 2026-10-09. Data is structure only — env values and anything secret-looking are stripped.

Similar rigs

kapadias/nonna

Your AI agent says done. Nonna makes it prove it: she runs your test suite before a coding agent can stop, and blocks pushes to main and secrets in files. Claude Code plugin, plus git hooks for Codex, Cursor, Copilot, Gemini and more

★ 0 project Claude CodeCodexCursor 27 skills 8 agents 11 hooks 5d ago

Fort Knox 3.9k tok ·

damilola-elegbede-org/claude-config

Configurations Files for Claude

★ 0 setup Claude Code 5 mcp 48 skills 7 agents 45 hooks 8d ago

YOLO Cowboy 35.2k tok ·

pruthvidarji1993/ai-automation-jira

AI-assisted dev workflow: Jira ticket in, reviewed PR out. Claude Code plugin + GitHub Copilot prompts with human approval gates (plan, diff).

★ 0 project Claude CodeCodex 9 skills 4 agents 1 hooks 1d ago

Pragmatist 2.0k tok ·

yoshpy-dev/ralph

Claude Code + Codex harness engineering — scaffold, upgrade, and run opinionated agent harnesses with deterministic hooks, evidence-backed review pipelines, and an autonomous multi-seat org runtime.

★ 6 project Claude CodeCodex 12 skills 10 agents 14 hooks today

YOLO Cowboy 4.7k tok ·