mtperessoni/ai-starter-kit
Make any repository AI-driven: PRD and TRD as the source of truth, a gate for every change, efficient subagents, related-tests-only workflow and an AI-readable structure, adapted to any stack by Claude Code skills.
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 mtperessoni/ai-starter-kit/kit/.claude ./rig-ai-starter-kit # 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 mtperessoni/ai-starter-kit/kit/.claude/skills/adr .claude/skills/adr $ npx degit mtperessoni/ai-starter-kit/kit/.claude/skills/docs-html .claude/skills/docs-html $ npx degit mtperessoni/ai-starter-kit/kit/.claude/skills/prd-create .claude/skills/prd-create $ npx degit mtperessoni/ai-starter-kit/kit/.claude/skills/prd-flow .claude/skills/prd-flow $ npx degit mtperessoni/ai-starter-kit/kit/.claude/skills/trd-create .claude/skills/trd-create
$ curl -fsSL --create-dirs -o .claude/agents/efficiency-auditor.md https://raw.githubusercontent.com/mtperessoni/ai-starter-kit/main/.claude/agents/efficiency-auditor.md $ curl -fsSL --create-dirs -o .claude/agents/prd-flow-docs.md https://raw.githubusercontent.com/mtperessoni/ai-starter-kit/main/kit/.claude/agents/prd-flow-docs.md $ curl -fsSL --create-dirs -o .claude/agents/prd-flow-executor.md https://raw.githubusercontent.com/mtperessoni/ai-starter-kit/main/kit/.claude/agents/prd-flow-executor.md $ curl -fsSL --create-dirs -o .claude/agents/prd-flow-recheck.md https://raw.githubusercontent.com/mtperessoni/ai-starter-kit/main/kit/.claude/agents/prd-flow-recheck.md $ curl -fsSL --create-dirs -o .claude/agents/prd-flow-reviewer.md https://raw.githubusercontent.com/mtperessoni/ai-starter-kit/main/kit/.claude/agents/prd-flow-reviewer.md $ curl -fsSL --create-dirs -o .claude/agents/prd-flow-surveyor.md https://raw.githubusercontent.com/mtperessoni/ai-starter-kit/main/kit/.claude/agents/prd-flow-surveyor.md
Merge into .claude/settings.json (project) or ~/.claude/settings.json (user). Hook commands reference scripts in the source repo — copy those too.
{
"hooks": {
"PreToolUse": [
{
"matcher": "*",
"hooks": [
{
"type": "command",
"command": "python scripts/telemetry_hook.py"
}
]
}
]
}
} Skills (5)
Subagents (6)
| efficiency-auditor model: opus | Runs the kit's eval rounds and computes efficiency from the metrics every time (tokens, cost per accept, wall time, output quality, errors), with subagent management and context flooding as the first |
| prd-flow-docs model: opus | prd-flow docs writer. The chief dispatches it to turn the user's answers into approved rules, to write the PRD (and the ADR a protected rule needs), the TRD Planned section and the plan, to fix a stal |
| prd-flow-executor model: sonnet | prd-flow executor. The chief dispatches one per task card, per fix (review findings or a routed failure) and once to close a delivery; it implements test first, commits its own work with the trailer a |
| prd-flow-recheck model: haiku | prd-flow scoped re-check. The chief dispatches it after a Critical or High finding was fixed; it confirms each previous finding against the fix commit only, findings only. |
| prd-flow-reviewer model: sonnet | prd-flow reviewer. The chief dispatches one at the end of each execution wave (and after the single task of a bug fix, implementation or refactor); the only reader of diffs, findings only, each routed |
| prd-flow-surveyor model: opus | prd-flow surveyor. The chief dispatches it first in every case and for a rule change in the middle of execution; it classifies, proves the rules against the code, sweeps conflicts and impact across ev |
Hooks (12)
| event | matcher | runs |
|---|---|---|
| SessionStart | * | python scripts/telemetry_hook.py |
| UserPromptSubmit | * | python scripts/telemetry_hook.py |
| PreToolUse | * | python scripts/telemetry_hook.py |
| PostToolUse | * | python scripts/telemetry_hook.py |
| PostToolUseFailure | * | python scripts/telemetry_hook.py |
| SubagentStart | * | python scripts/telemetry_hook.py |
| SubagentStop | * | python scripts/telemetry_hook.py |
| Notification | * | python scripts/telemetry_hook.py |
| PreCompact | * | python scripts/telemetry_hook.py |
| PostCompact | * | python scripts/telemetry_hook.py |
| Stop | * | python scripts/telemetry_hook.py |
| SessionEnd | * | python scripts/telemetry_hook.py |
Permissions
deny (0)
—
ask (0)
—
allow (16)
Bash(scripts/gates.sh:*)
Bash(bash scripts/gates.sh:*)
Bash(python scripts/ratchet.py:*)
Bash(python scripts/related_tests.py:*)
Bash(python scripts/new_failures.py:*)
Bash(python scripts/hotspots.py:*)
Bash(python scripts/contract_drift.py:*)
Bash(python .claude/skills/prd-flow/scripts/gate.py:*)
Bash(python .claude/skills/prd-flow/scripts/build_prd_html.py:*)
Bash(python .claude/skills/prd-flow/scripts/build_trd_html.py:*)
Bash(python .claude/skills/prd-flow/scripts/promote.py:*)
Bash(git status:*)
Bash(git diff:*)
Bash(git --no-pager diff:*)
Bash(git log:*)
Bash(git show:*)
Similar rigs
rness-dev/rness
One source of truth for every AI coding agent. The standards, decisions, specs and plans of your organization, written into the AGENTS.md of every repository. Claude Code, Codex, Cursor and GitHub Copilot already read it.
Automator 5.6k tok ·
max-ranger/dotfiles
Personal dotfiles: Claude Code config, skills, hooks, Brewfile
Pragmatist 1.0k tok ·
theNickNance/claude-code-setup
—
Pragmatist 3.5k tok ·
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.
Pragmatist 747 tok ·