Conciseness

The body is largely lean — no explanations of concepts Claude already knows (no 'what is NIH' or library tutorials), and the intake questions, curl templates, and tables are all operational content. Minor trimmable material remains: the hardcoded submission receipt dates ('Feb 5, Jun 5, Oct 5' etc.) are time-sensitive and not in a deprecated/verification section, and the build-provenance footer ('Source spec ... gitignored, not present', 'PR #657 audit', 'Path B (direct conversion)') is maintainer noise irrelevant to the executing agent. This fits anchor 4 ('Efficient; minor instances ... that could be trimmed') better than anchor 5, and better than anchor 3, which requires more substantial unnecessary explanation.

Actionability

Guidance is fully executable: complete copy-paste curl POST commands with full JSON payloads for both narrow (AND) and broad (OR) RePORTER searches, exact script invocations with flags ('python scripts/mechanism_matcher.py --career-stage "early_career" --prelim-data "pilot" ...'), literal intake question scripts with option lists, the NOSI URL pattern, and a runnable docx validation command ('python3 -c "import zipfile,sys; zipfile.ZipFile(sys.argv[1]).testzip()"'). This matches the score-5 anchor (copy-paste ready, common cases covered) rather than anchor 4, which allows minor gaps — placeholders like <research idea> are the only substitutions needed.

Workflow Clarity

The workflow is clearly sequenced Phase 1 (intake with a hard stop condition: 'After Q6, commit and start Phase 2A. Never re-open intake') → Phase 2A/2B → Phase 3 → Phase 4, with explicit validation in Phase 4 (zip-integrity check plus confirming required sections are present), feedback loops (retry once after 3s; after 3 consecutive failures stop and alert; 'Consensus returns 0 for a facet → surface explicitly'), and a checklist-style Anti-Patterns section. This matches the score-5 anchor (explicit validation steps, error-recovery feedback loops, checklists) rather than anchor 4, which permits minor validation gaps.

Progressive Disclosure

The body is a genuine overview that splits detail into three one-level-deep, well-signaled reference files (references/reporter_post_patterns.md, references/nih_mechanism_matching.md, references/docx_9_sections.md — all verified to exist), linked both inline and in a dedicated '## References' section, plus a '## Tooling' table for the three scripts. The full DOCX spec and mechanism matrix live in references while only summaries stay inline. This matches the score-5 anchor (clear overview, well-signaled one-level-deep references, appropriate split) rather than anchor 4, which expects minor organization gaps.

Specificity

The description lists multiple concrete actions with comprehensive coverage: 'Grill-me intake (research idea + career stage + preliminary data + environment + submission posture + known institute targets)', 'Runs a 5-facet Consensus positioning analysis', 'maps the research to the right NIH institutes and study sections via RePORTER', 'finds NOSIs and funded overlap', 'produces an editable Word document (.docx) with budget/scope-aware mechanism recommendations, submission timelines, and a mandatory program officer recommendation'. This matches the score-5 anchor ('Lists multiple specific concrete actions; comprehensive coverage') and exceeds anchor 4, which is for coverage with minor gaps — no capability area of the workflow is missing.

Completeness

Both 'what' and 'when' are answered explicitly and concretely. The 'what' is a detailed capability list (intake, positioning analysis, institute mapping, NOSI discovery, docx deliverable), and the 'when' is a full 'Use when...' clause with five concrete trigger phrases. This matches the score-5 anchor exactly ('Clearly and explicitly answers both what AND when with concrete trigger phrases'); it is not anchor 4, where the 'when' is present but could be more explicit or specific.

Trigger Term Quality

Trigger coverage is comprehensive with natural user phrasings and synonyms: 'grants for [topic]', 'find grants for my research idea', 'what grants match my research', 'help me find NIH funding', 'grant opportunities for my research', plus the general 'research funding' and 'any grant-related request'. This matches the score-5 anchor (comprehensive natural terms including synonyms) rather than anchor 4 ('a few natural terms missing') because the example phrases span the realistic ways a user would ask for grant discovery; terms like 'grant writing' are absent but describe a different task outside this skill's purpose.

Distinctiveness Conflict Risk

The skill occupies a clear NIH-grant-discovery niche with distinct triggers and an explicit scope boundary: 'NIH-only scope — non-NIH funders (PCORI, DOD CDMRP, VA, foundations) are out of scope and flagged at intake.' This actively prevents mis-triggering on general funding or literature-search requests, matching the score-5 anchor (clear niche with distinct triggers; minimal conflict risk) rather than anchor 4, which allows minor overlap risk with closely related skills.