docs(plugins): normalize existing cross-skill chains to explicit Skill-tool phrasing (invocation-mode doctrine sweep)

[kyle-sexton]

Context

Lane 6 of the AI Hero course vetting adopted the cross-skill invocation-phrasing doctrine (docs/conventions/invocation-mode/README.md "Cross-skill invocation phrasing"): skill bodies that chain to another skill name the mechanism explicitly ("invoke /plugin:skill via the Skill tool"), never bare /name prose. PR #3001 review (Codex P2) correctly observed the fleet's existing practice is MIXED — e.g. plugins/knowledge/skills/map-corpus/SKILL.md hands work to /knowledge:docpage-digest and /planning:interview bare, as does plugins/knowledge/skills/docpage-digest/SKILL.md — so the doctrine is scoped to new and edited skill text until this sweep lands.

Proposed work

• Enumerate operative cross-skill chains in plugins/**/SKILL.md (and agent/reference docs that instruct invocation) that use bare /name prose, distinguishing operative instructions from mere mentions/examples.
• Rewrite operative chains to the explicit Skill-tool phrasing; leave prose mentions alone.
• Consider a skill-quality:check criterion or lint pattern to keep new text conformant (coordinate with the rubric-enforcement scope of issue skill-quality/fleet: enforce the invocation-mode rubric — explicit disable-model-invocation on 17 skills, check criterion, skill-authoring pointer #2968 — same doctrine home, different axis).

Acceptance criteria

• Operative bare-/name cross-skill chains rewritten fleet-wide (or an explicit per-case keep rationale recorded)
• The doctrine paragraph's "new and edited text" scoping note updated to reflect the completed sweep
• Enforcement decision recorded (criterion/lint added, or explicitly declined with reason)

References

• docs/conventions/invocation-mode/README.md ("Cross-skill invocation phrasing")
• docs/upstream/aihero-course.md (term-adoption section — the adoption decision and its upstream provenance)
• PR docs(upstream): AI Hero course lanes 3–6 — decision records and consolidated provenance #3001 (the Codex review finding); issue skill-quality/fleet: enforce the invocation-mode rubric — explicit disable-model-invocation on 17 skills, check criterion, skill-authoring pointer #2968 (sibling rubric enforcement)

Metadata

Field	Value
Category	unspecified
Area	docs, skill-quality
Ecosystem	unspecified