Conciseness

The body is lean and compressed — no explanations of concepts Claude already knows, and jargon like 'Goldilocks' and 'OMTM' is used without padding. However, the Rules section restates workflow content ('Indicators without current values are flagged n/a, not invented' repeats step 3; 'One card, at most one learning, board regenerated' repeats steps 6–7), so it sits at 'efficient; minor instances that could be trimmed' rather than every token earning its place.

Actionability

Highly concrete for a judgment-based skill: a fill-in hypothesis template ('if [user/stakeholder] attains [benefit] via [change], we expect [business impact]'), an explicit risk-matching rule (low → just-do-it; medium → ship & measure; high → discovery/testing), named fields ('outcome_hypothesis', 'risk_level', 'derisking_approach'), and a concrete bad example ('12 features shipped'). Not 5 because 'contract format' for learnings and 'board regenerated' are invoked without a template or pointer, leaving steps 6–7 not fully executable from this file alone.

Workflow Clarity

Seven clearly sequenced steps (read cold → rewrite → indicators → assumptions → derisking fit → record → capture) close with an explicit Quality Gates checklist, and failure paths are handled ('or the brief says why not, and what's missing'; mismatch 'escalated as a candidate move'). This matches the anchor of a clear sequence with explicit validation steps and a checklist, and the single-card scope means no batch/destructive validation cap applies.

Progressive Disclosure

Good structure with well-organized sections and a single clearly signaled one-level-deep reference ('Ladder + Goldilocks/OMTM: skills/sniff-test/references/derisking-and-indicators.md (this repo)'). Docked from 5 because that path lives outside the skill's own bundle (no references/ directory ships with it) and in-body terms ('watermelon risk', 'seek-alpha', 'contract format') have no navigable definition — minor organization gaps.

Specificity

Lists four concrete actions — 'rewrite activity-framed cards into outcome hypotheses with leading indicators, sharpen riskiest assumptions with evidence/opinion tags, fix derisking-approach mismatches, and capture at most one durable learning' — which comprehensively cover the body's workflow. Matches the multiple-specific-actions/comprehensive-coverage anchor; the abstract lead-in ('Improve the quality, outcome orientation, and derisking') is immediately made concrete.

Completeness

Both halves are explicit: what = the four enumerated rewriting/sharpening/capture actions; when = 'Use on cards flagged activity-oriented, thin on evidence, missing indicators, or carrying committed money with open risks.' Concrete trigger phrases are present, matching the clearly-and-explicitly-both anchor.

Trigger Term Quality

'Use on cards flagged activity-oriented, thin on evidence, missing indicators, or carrying committed money with open risks' provides natural trigger phrases a user of this system would actually say. Falls just short of comprehensive: common variations such as 'outcome-oriented', 'OKR', 'hypothesis', or 'steering' are absent — good keyword coverage with a few natural terms missing.

Distinctiveness Conflict Risk

'The loop-2 skill of the portfolio-pdlc family' positions it within its family, and the narrow, flag-based triggers ('carrying committed money with open risks', 'missing indicators') carve a clear niche with minimal conflict risk — they would not naturally fire for sibling skills or generic planning skills. Not 4 because the overlap with closely related family skills is explicitly resolved by the loop-number positioning.