100%

Goal

Rewrite /intel-package SKILL.md as a thin, script-first body (per script-first.md §1 — directives + action-consumption, no orchestration prose); doc sweep across intelligence.md, the query-intel surfaces doc, and CLAUDE.md pointers to reflect the dossier genre + unified heal.

Implementation Sketch

  • SKILL.md: cite the runtime invocation python -m scripts.intel_package_runtime <subcommand>; cite the tier-selection SSOT module scripts/intel_package_runtime/ (delivered by s-aa802633) for target-kind tier selection — do NOT carry/encode the tier-selection map or logic in the skill body; @-include the consumed rules; remove any orchestration narrative.
  • Docs: update intelligence.md (Intelligence Package section), the query-intel surfaces doc, CLAUDE.md §Intelligence Graph pointer to the new genre + heal.

Spec References

  • .claude/skills/intel-package/SKILL.md
  • .claude/rules/script-first.md §1, skill-template.md §2
  • scripts/intel_package_runtime/ (tier-selection SSOT; target_kinds.py five-kind set)
  • intel-package-rebuild-notes.md R1 (genre inversion), R3 (target-kind tiers), R10 (per-section generation), R13 (unified heal), R5 (DIG DEEPER blocks) — the dossier-genre requirements the doc sweep must reflect

Work Items

  • Thin script-first SKILL.md rewrite: runtime-invocation citation, tier-selection SSOT-module citation (no encoded map), rule @-includes, subcommand citations.
  • Doc sweep: intelligence.md, query-intel surfaces, CLAUDE.md pointers.

Verification

  • sc-0010 is FLAGGED-FOR-EXECUTION: tighten its probe at execution phase via amend-success-criterion from the generic pytest bucket to named pins — scripts/skill_audit/script_first_audit.py --skill intel-package + scripts/prose-lint.py on the rewritten SKILL.md + doc-pointer assertions that intelligence.md, the query-intel surfaces doc, and CLAUDE.md §Intelligence Graph reference the dossier genre + unified heal (all three doc-sweep targets named in the Goal + Implementation Sketch + w-afdf36d0). plan.json is route-state (script-only); this flag is a body note matching the sibling sc-0008/sc-0009 pattern, not a plan.json edit.

HISTORY

  • 2026-06-13 — /review-plan Step 5 editor edits: Reworded Implementation Sketch + w-7c17c932 so the SKILL.md CITES the runtime invocation + tier-selection SSOT module (scripts/intel_package_runtime/, delivered by s-aa802633) rather than encoding the tier-selection map in the skill body (cures bs-tier-map-in-skill-body / cc-script-first-thin-body — avoids STRUCTURE:script-first-violation). Broadened Spec References from R5 alone to R1+R3+R10+R13+R5 covering the dossier-genre scope the doc sweep conveys (bs-spec-ref-groundedness). Added Verification block flagging sc-0010 for execution-phase probe tightening (bs-sc-0010-probe-mismatch). depends_on:[] reviewed and left as-is — defensible per audit (predecessor runtime sections s-fa1f33bc + s-9a707429 complete, lighter-rigor doc section, no DAG contradiction).