Conciseness

Mostly efficient and assumes Claude's competence, but the narrative intro and 'How it differs from a quick research router' section overlap with 'Core mechanisms', adding minor redundancy that could be trimmed.

Actionability

Concrete guidance throughout — specific file names (plan.md, sources/NN.md, refresh_targets.md), explicit counts (≥3 independent sources, 2–4 hypotheses, 4 self-critique questions) and named genres — but a few referenced artifacts (the 4 self-critique questions, opposition queries) are not enumerated.

Workflow Clarity

The 9-phase pipeline is clearly sequenced with validation gates (triangulation ≥3 sources, Phase 6.5 'Verify' citation check, adversarial pass, stop-criteria), though the verify step is described only as 'lightweight' without a feedback loop.

Progressive Disclosure

Well-organized into clear sections (pipeline, core mechanisms, output structure, when-to-use, anti-patterns, cross-references) with no nested references; it is a self-contained single file, so structure is good though slightly beyond the 'simple skill' threshold.

Specificity

Lists multiple concrete actions — 'fan-out web search across many channels, parallel sub-agents, source triangulation... adversarial review pass, and every source saved to its own file with verbatim quotes' — giving comprehensive coverage of capabilities.

Completeness

Explicitly answers both 'what' (the multi-source investigation pipeline) and 'when' via a concrete 'Use when a low-quality answer is expensive' clause plus negative 'NOT for...' triggers.

Trigger Term Quality

Good natural trigger phrases ('strategy work, comparing N products/methods/markets, validating a hypothesis with external data, mapping how a field works') but a few common variations (e.g. 'competitive analysis', 'due diligence') are absent.

Distinctiveness Conflict Risk

Carves a clear niche and explicitly disambiguates from siblings via 'NOT for... (use competitive-teardown)' and 'use the research router instead', minimizing conflict risk.