Author a governed spec change

1. Read AGENTS.md, spec/README.md, the schema, relevant engine validation, and the complete governing decision or accepted-direction record.
2. Require a pui-trace map and current-user or standing authorization for the fixed identity, owner, lifecycle intent, semantic boundary, compatibility, and evidence scope.
3. Search existing IDs and relations again before creating anything.
4. Model one coherent semantic slice. Write addressable criteria, precise directional relations, version bounds, sources, open questions, and revisions where semantics changed. Record lifecycleRationale for new ordinary entities or lifecycle changes, classify canonical openQuestions[].blocks targets, and validate the changed authoring scope with pnpm check:spec-authoring -- --base <base-sha> as defined in spec/README.md.
5. Add or update T-* evidence and real executable implementation paths for normative behavior.
6. Change generated projections through their generators.
7. Record spec graph, catalog, projection, focused evidence, and proportional type checks for a later pui-validate transition.

Route an actually undecided schema design, ownership, public surface, compatibility, lifecycle promotion, or evidence meaning to the smallest product-direction decision. Do not create inventory placeholders or infer a matrix from files, dependencies, packages, or names.

Explicit handoff

Do not load or execute another skill. Return exactly one handoff conforming to internal/agent-operations/schemas/skill-handoff.schema.json. Carry required prior artifacts by reference, include every artifact this leaf produces according to skills.yaml, and set nextSkillId to one eligible registered leaf or null.

Communicate with the user in the user's current language. Preserve canonical identifiers and governed localization.