agentleFS
t1djanit1djani/servo/skills/shape-spec/SKILL.md

shape-spec

Skill1 starsChanged 5 months agoMITView on GitHub

Write a spec that carries its own guardrails — it embeds the invariants it must not break and names the oracles it will be judged against — then gate it. Use after the design forks are decided and before writing a plan.

The file

Files
SKILL.md

shape-spec

Turn a decided design into a spec that the rest of the flow can trust. A good servo spec is not just a description; it carries the means to check itself. Two moves make that happen.

The two moves

  • Embed the invariants inline (constrain). Whatever the work must never break — pulled from the invariants oracle and the relevant experts — is written into the spec as an explicit list, not left implicit. A constraint that is physically present in the document cannot be silently forgotten downstream.
  • Name the oracles (point, do not freeze). The spec declares which sources it will be judged against — scope, invariants, acceptance, prior-art — by reference, so they stay fresh. (A plan, later, will snapshot its acceptance target instead; a spec evolves, so it points.)

Procedure

  1. Start from the context brief (gather-context) and the decided forks (expert-panel). Do not re-open settled decisions here.
  2. Write the spec. State the goal, the approach, and the scope — including what is explicitly out of scope.
  3. Add an Invariants section. List, inline, what must not break, each tagged with the oracle/expert it came from.
  4. Add an Oracles section. Name the manifest keys this spec will be gated against.
  5. Gate the spec. Run servo-gate on the spec, once per relevant oracle: against invariants (does it violate any?), scope (does it match the intent, no creep?), prior-art (does it ignore or contradict what exists?), footguns (does it walk into a known trap?). Fix what comes back FIX/STOP; do not proceed on an open finding.

Rule

The spec is the last place divergence is allowed. After it, the plan is pure convergence. If a real fork is still open, go back to expert-panel — do not bury the choice in the spec for the plan to trip over.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.