skills.sh
AvailableUniversal skill installation, including Claude Code and Codex.
npx skills add https://github.com/ulpi-io/skills-autonomous-engineering --skill auto-specauto-spec
Turns a raw request into a repo-grounded, testable specification with explicit non-goals.
WORKFLOW / 05/auto-spec <what to build — the feature/product/change request>
- 01
Intake the request and stop on build-changing ambiguity.
- 02
Recon the repository, domain, interfaces, constraints, and existing behavior.
- 03
Draft objectives, observable behavior, acceptance criteria, non-goals, assumptions, and risks.
- 04
Run a completeness critic until material gaps are gone or progress stalls.
- 05
Write the finalized specification under .ulpi/spec/.
▸ operating envelope
Use it deliberately.
The shortest useful definition of where this skill helps—and where it should stay out of the way.
Use when
- Starting a new feature, product, or significant change.
- A broad request needs repository recon and measurable acceptance criteria.
- Planning needs a written contract for what will and will not be built.
Do not use when
- The product decision is still too ambiguous to specify; ask the blocking questions first.
- You need implementation tasks rather than requirements; use auto-plan.
- You are ready to write code; use auto-build after an approved plan.
▸ output contract
What must be true when it stops.
Produces
- .ulpi/spec/<name>.md
- Testable acceptance criteria and explicit non-goals
- Assumption and open-question ledger
- Completeness-critic convergence report
Guarantees
- Every requirement is grounded in observed reality or labeled as an assumption.
- Every acceptance criterion names an observable, testable result.
- Scope includes explicit non-goals.
- Material ambiguity is surfaced rather than silently guessed.
- A stalled critic returns the open gaps instead of a false complete verdict.
Honest failure states
- blocked_ambiguity — a product decision changes what would be built
- stalled — the critic cannot close the remaining material gaps
- incomplete — criteria remain untestable or ungrounded
▸ enforcement
Know which claims are executable.
Scripts are linked when this skill owns deterministic machinery. The remaining rules are explicit operating contracts and must not be marketed as hooks.
Procedural contract
This skill composes platform tools and verification evidence; it does not claim a dedicated guard script.
▸ install and invoke
One skill, three honest surfaces.
Claude Code plugin
AvailableInstall the full plugin, then invoke /auto-spec or let routing select it from context.
Codex plugin
Work in progressA work in progress on the codex-native-plugin branch — build it, start a new session, then invoke $autonomous-engineering:auto-spec.
▸ composes with