本文へ移動
cccskills
無料GitHub で公開

project-constitution

Author the project's non-negotiable engineering rules — the constitution that every feature spec, plan, and implementation must satisfy. Load when the user asks to write a constitution, define project rules, set engineering invariants, write project standards, define non-negotiables, capture agent guardrails as policy, or when spec-driven-development orchestrator routes here. Also triggers on "project constitution", "engineering invariants", "non-negotiable rules", "project policy", "set our standards", "/constitution". Output: docs/constitution.md. The third strategic artifact alongside AGENTS.md (agent behavior) and product-soul (strategy).

インストール方法を見る

含まれるファイル(2)

  • SKILL.md7.2 KB
  • references/examples.md2.3 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Project Constitution

You are a Principal Engineer. You write the project's irreducible engineering invariants — the rules that NEVER bend, regardless of feature, sprint, or PR. Specs reference the constitution. Plans cannot violate it. Implementations are rejected if they break it.

Hard Rules

Never write a constitution longer than ~120 lines — it must fit in every agent's context. Never write rules without IDs — every rule needs C-N so specs and plans can cite them. Never write aspirational rules — every rule must be enforceable in code review or CI. Never duplicate AGENTS.md (agent behavior) or product-soul (strategy) — constitution is engineering invariants only. Never amend silently — every change bumps the version and gets an Amendment line.


Workflow

Step 1 — Check existing context

Read in priority order:

  1. Existing docs/constitution.md (if updating — preserve IDs).
  2. AGENTS.md (don't duplicate agent behavior rules).
  3. docs/product-soul.md (don't duplicate strategy).
  4. package.json / pyproject.toml / Cargo.toml / go.mod / Makefile (extract de-facto current standards: lint, test, type-check, coverage thresholds).

Step 2 — Interview (max 5 questions, one at a time)

For each constitution category below, ask only what cannot be inferred:

  1. Testing — required coverage, types of tests, what's untested-allowed
  2. Security/Privacy — credentials, PII, dependency audit posture, threat model
  3. Performance — latency budgets, bundle size limits, p95/p99 targets
  4. Accessibility — WCAG level, supported languages (skip if no UI)
  5. Dependencies — license rules, version pinning, vendor allowlist
  6. Observability — logs, metrics, traces, error tracking required
  7. Migration/Rollback — DB migrations, feature flags, blast radius
  8. Documentation — what must be documented before merge

Skip categories that don't apply.

Step 3 — Write the constitution

Use the schema below. Each rule:

  • Has a stable ID (C-1.1)
  • Is normative (MUST / MUST NOT / SHOULD)
  • Is enforceable (a reviewer or linter can detect violations)
  • Has at most a 1-line rationale

Step 4 — Self-review

  • Each rule has a unique ID
  • No "should consider" / "try to" / "where possible" — vague = unenforceable
  • No rule duplicates AGENTS.md or product-soul
  • Total length < 120 lines
  • Versioned (header has Version + Date)

Step 5 — Save and notify

Save to docs/constitution.md. Append to docs/skill-outputs/SKILL-OUTPUTS.md:

| YYYY-MM-DD HH:MM | project-constitution | docs/constitution.md | Constitution v<N> |

Tell the user:

"Constitution saved to docs/constitution.md (v<N>). Every feature-spec must reference this version. Re-run when rules need to change — never amend silently."


Output Format

# Project Constitution
Version: 1 | Date: YYYY-MM-DD | Status: Active

## C-1 Testing
- C-1.1 Every PR MUST include unit tests for new logic. Rationale: regression protection.
- C-1.2 ...

## C-2 Security/Privacy
- C-2.1 Credentials MUST come from env vars or secret manager — never committed.
- C-2.2 ...

## C-3 Performance
- C-3.1 ...

## C-4 Accessibility (only if UI)
- C-4.1 ...

## C-5 Dependencies
- C-5.1 New runtime deps require ADR if they add > 100KB to bundle.

## C-6 Observability
- C-6.1 Every API endpoint MUST emit a request-id correlated log line.

## C-7 Migration & Rollback
- C-7.1 Every DB migration MUST be reversible OR shipped behind a feature flag.

## C-8 Documentation
- C-8.1 Public API changes MUST update docs in same PR.

## Amendments
- 2026-05-02: Initial draft.

Gotchas

  • Constitution rules are NOT preferences — they bend for nobody. If a team consistently waives a rule, remove it instead of letting it rot.
  • Each rule must point at a single observable behavior. "Code should be clean" is not a rule. "Functions MUST be < 50 lines OR have a # noqa: complexity comment" is.
  • Version bumps: any rule change is a major bump. Add to Amendments. Specs cite the version they were written against (e.g. constitution: docs/constitution.md@2).
  • AGENTS.md vs constitution: AGENTS.md tells the agent HOW to work. Constitution tells the project WHAT must always be true. Don't confuse them.
  • A waiver mechanism exists in feature-spec (a ## Constitution Waivers section). Don't try to encode waivers here.

Example

<examples> <example> <input>Set up a constitution for our SaaS app — Next.js + Postgres + Vercel.</input> <output> [Interview: 4 questions — testing posture (90% on services, no UI snapshot), perf budget (LCP < 2s), GDPR (yes, EU users), DB migration policy (always reversible)]

Wrote 8 categories, 19 rules, 94 lines. Top rules:

  • C-1.1 Service layer code MUST have ≥ 90% coverage (Vitest)
  • C-2.2 PII fields MUST be marked in Prisma schema with @encrypted
  • C-3.1 LCP MUST be < 2s on p75 (Vercel Analytics)
  • C-7.1 Every Prisma migration MUST be backwards-compatible across one deploy

Saved to docs/constitution.md v1. Every feature-spec must now reference constitution: docs/constitution.md@1. </output> </example> </examples>


Common Rationalizations

ExcuseReality
Constitution is boilerplateC-N rules must be project-specific and testable.
Skip version bumpAmendments need version + date for spec linkage.
Copy from template onlyInterview user for real non-negotiables.
One page is enoughDepth on gates beats vague values.

Verification

  • Version and date in constitution header
  • C-N items are observable and enforceable
  • Linked from feature-spec workflow
  • Amendment process documented

Red Flags

  • Rule stated as preference — routinely waived in practice
  • Unenforceable rule like code should be clean
  • Constitution changed without version bump and amendment
  • Spec approved against outdated constitution version

Prune Log

Last pruned: 2026-07-04

  • No changes — citation audit passed; content current (improve-skills full pass 2026-07-04)

Impact Report

Constitution complete: v<N>
Categories: <N> populated, <M> skipped (with reason)
Rules: <N> total
Lines: <N>/120
Saved: docs/constitution.md
Logged: docs/skill-outputs/SKILL-OUTPUTS.md
Cited by: feature-spec, implementation-plan, spec-crosscheck

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Put on the adversarial hat and systematically attack any document, plan, strategy, or idea to expose its weakest points before commitment. Structured devil's advocate with red team rigour — not pessimism, but evidence-based critique across three phases: diagnostic (are claims accurate?), creative (is the problem artificially constrained?), challenge (are solutions robust?). Load when the user asks to stress test a document, red team this plan, poke holes in this, devil's advocate this, challenge my assumptions, or when product-soul, brainstorming, prd-writing, or inversion calls for adversarial review. Also triggers on "what am I missing", "what could kill this", "find the flaws", or "critique this rigorously".

日本語の概要は準備中です。原文の説明を表示しています。

dvy1987/agent-loom32026年8月8日 更新

Design execution structure for decomposed processes: single agent or multi-agent topology. Load when user says "design an agent for this", "what agent structure do I need", "architect this", "should this be multi-agent", "what's the right execution structure", "agent topology", "how should agents be organized". Takes process-decomposer output as primary input. If triggered directly without a process entry, calls process-decomposer first.

日本語の概要は準備中です。原文の説明を表示しています。

dvy1987/agent-loom32026年8月8日 更新

Internal skill. Called by setup-evaluation after a PASS. Launches agents from a validated architecture spec using Claude Code / Ampcode native parallelism (Task tool). Does NOT generate scripts or SDK code — it outputs structured spawn instructions that the platform executes natively. Never invoked directly by the user. Never launches without a setup-evaluation PASS.

日本語の概要は準備中です。原文の説明を表示しています。

dvy1987/agent-loom32026年8月8日 更新

Sync library skills from an agent-loom upstream repo into this project's .agents/skills while preserving project-local and forked skills. Load when the user asks to sync agent-loom, update skills from upstream, rsync from ../agent-loom, pull new library skills, upgrade installed skills, or refresh the .agents folder without losing custom project skills. Also triggers on "sync skills from agent-loom", "update my agent skills", "pull skill library updates", or "merge agent-loom improvements into this repo".

日本語の概要は準備中です。原文の説明を表示しています。

dvy1987/agent-loom32026年8月8日 更新

Instrument a shipped product's AI agents with tracing and observability so you can see what they did, why outputs happened, and what each run cost. Plain-language primer plus free-tier-first backend selection (Langfuse, Phoenix, LangSmith, Braintrust) and OpenTelemetry/OpenInference instrumentation. Load when the user asks to add observability, add tracing, instrument my agents, see what my agent is doing in production, set up Langfuse or Phoenix or LangSmith, debug why my agent gave a bad answer, or track LLM cost per request. Also fires when agent-system-architecture or setup-evaluation requires an observability plan for an agent-chain product. NOT for tracing the coding agent itself — that is run-trace. Precondition for runtime-learning-loop.

日本語の概要は準備中です。原文の説明を表示しています。

dvy1987/agent-loom32026年8月8日 更新

Run a structured retrospective after development-phase runs of your product's agents — interview the owner in plain language about what went well and poorly, draft ranked improvement hypotheses, then design and run small n=1/n=2 experiments with pre-declared success criteria, guardrails, stop conditions, and a cost/ROI kill-switch. Load when the user says how did that run go, retro this run, the agent output was bad, what should we improve, draft hypotheses, run a small experiment, or after repeated dev runs of an agentic system produce uneven quality. Priority: output quality over performance over cost, each with diminishing-returns stops. NOT a product A/B test (experimentation), NOT coding-agent harness repair (harness-evolution), NOT production-scale learning (runtime-learning-loop).

日本語の概要は準備中です。原文の説明を表示しています。

dvy1987/agent-loom32026年8月8日 更新

dvy1987 のスキルをすべて見る

このスキルの問題を報告する