Conciseness

The body is dense and operational — phase structure, exact query templates, full curl payloads, error-handling and anti-pattern tables — with almost no explanation of concepts Claude already knows. Minor redundancy keeps it from a 5: the Agent Integrity Rules, Error Handling table, and Anti-Patterns list repeat the same retry/stop/no-fabrication rules three ways, the NIH-only scope statement appears in both a blockquote and the frontmatter, and phrases like 'This is the single most valuable advice for any applicant' are padding.

Actionability

Mostly executable guidance: complete RePORTER curl POST examples with full JSON payloads, script invocations with flags and expected output ('python scripts/fiscal_year_calculator.py --output json' with a sample return), exact Consensus query strings, a concrete output filename convention, and a runnable zip-integrity check. The gap is that DOCX generation — the core deliverable — has no code inline and delegates entirely to 'references/docx_9_sections.md', which is not present in the bundle, so that step cannot actually be executed from the provided material.

Workflow Clarity

Clear four-phase sequence with a hard stop condition ('After Q6, commit and start Phase 2A. Never re-open intake'), per-tool feedback loops (wait 3s, retry once, escalate after 3 consecutive failures), and a final validation step (zip integrity test plus section-presence confirmation). It falls short of 5 because the DOCX validation is post-hoc with no failure branch — nothing says what to do if testzip() fails or a section is missing — whereas the level-5 anchor requires validate → fix → re-validate loops throughout.

Progressive Disclosure

The structure is right in intent — a References section listing three one-level-deep reference files plus three scripts — but scored against the actual bundle, none of the six referenced files ('references/reporter_post_patterns.md', 'references/nih_mechanism_matching.md', 'references/docx_9_sections.md', and the three scripts/*.py) exist, making every link dead. Additionally, the RePORTER curl templates are fully inlined despite the reference file claiming to hold them, so content is duplicated rather than split. This fits the level-3 anchor — some structure, references present, but content that should be separate is inline — not 4, where references would be real and mostly well-placed.

Specificity

The description lists multiple concrete actions with comprehensive coverage: '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', and 'produces an editable Word document (.docx) with budget/scope-aware mechanism recommendations, submission timelines, and a mandatory program officer recommendation'. Each capability is named specifically, matching the level-5 anchor's multiple specific concrete actions.

Completeness

It explicitly answers both questions: the 'what' is a full capability inventory (positioning analysis, institute mapping, NOSI discovery, DOCX output with mechanism recommendations and program officer recommendation), and the 'when' is an explicit trigger clause ('Use when the user asks about research funding or makes any grant-related request') with five concrete example phrases. This matches the level-5 anchor exactly.

Trigger Term Quality

Good keyword coverage with natural user phrasings ('research funding', 'find grants for my research idea', 'what grants match my research', 'help me find NIH funding', 'grant opportunities for my research'), but a few natural terms users would say are missing (e.g., 'funding opportunities', 'grant application', 'R01', 'apply for funding'). This fits the level-4 anchor — good coverage, a few natural terms missing — rather than 5, which requires comprehensive synonym coverage, and clearly above 3, since common variations are present.

Distinctiveness Conflict Risk

It occupies a clear niche (NIH funding strategy for clinical researchers) with distinct triggers, and the explicit boundary statement — 'NIH-only scope — non-NIH funders (PCORI, DOD CDMRP, VA, foundations) are out of scope and flagged at intake' — further reduces collision risk with adjacent literature-review or general-research skills. Minimal conflict risk, matching the level-5 anchor.