Conciseness

The body is mostly efficient — the routing table, forcing questions, and script commands all earn their tokens — but the same rules are stated three times: the Operating Prompt section ('Lead with the strongest counterargument', 'No disclaimers'), Hard Rules items 3–5 and 7 ('Counterargument first', 'Confidence levels mandatory', 'No sycophancy, no disclaimers'), and the Anti-Patterns list ('Validating the premise...', 'Citing an Andreessen quote without a confidence level...'). The plugin-comparison paragraph ('It is the Andreessen-lens counterpart to a founder-operating-system plugin') is decorative. This matches the level-3 anchor 'mostly efficient but includes some unnecessary explanation or could be tightened'; it is not level 2 because there is no explanation of concepts Claude already knows and no padding filler.

Actionability

Fully executable guidance throughout: a routing table mapping user intents to scripts, six forcing questions each with a recommended answer, exact copy-paste commands with realistic arguments ('python scripts/market_first_evaluator.py --size 8 --growth 7 --timing 9 --pull 8 --team 6 --product 5'), a fixed verdict vocabulary (BUILD-POUR-FUEL / MARKET-FIRST-DERISK / KILL-OR-REPICK-MARKET), and a worked delivery format. This matches the level-5 anchor: copy-paste ready commands covering the common cases.

Workflow Clarity

A clear four-step sequence — detect and route, forcing-question interrogation one at a time ('Do not batch them — make the user commit to each'), deterministic tool scoring ('the verdict isn't vibes'), then verdict delivery with confidence levels. This is not a destructive/batch skill so the validation cap does not apply, and the script-score gate functions as a checkpoint. It falls short of level 5 only because there is no explicit handling for edge states — e.g., what to do when inputs to a scoring script are unknown, or when the user refuses to answer a forcing question — leaving minor validation gaps per the level-4 anchor.

Progressive Disclosure

The body is a clean overview with well-signaled, one-level-deep references: a References section listing each of the 4 reference files with a one-line content description, an Assets section with 5 named templates/examples, and a Tooling table for the 3 scripts. Core workflow content (routing, forcing questions, commands) stays inline while deep material (canon sources, worksheets, worked examples) is pushed out, matching the level-5 anchor. Note: no bundle files were present in the reviewed directory to verify the referenced paths exist, so this score reflects the body's structure as written.

Specificity

The description lists multiple concrete capabilities: 'pressure-tests ideas, ventures, features, and career bets through Andreessen's actual frameworks', 'leads with the strongest counterargument, never validates premises, uses explicit confidence levels', and 'provides the 3x5-card + Anti-Todo personal productivity routine'. These are specific, distinct actions with comprehensive coverage of the skill's three functions (verdicts, PMF checks, daily productivity). It is not level 4 because no meaningful capability is left uncovered; not below 5 because nothing here is vague or abstract.

Completeness

Both 'what' (market-first pressure-testing of ventures via specific frameworks, PMF evaluation, 3x5-card/Anti-Todo routine, anti-sycophancy verdict delivery) and 'when' are explicitly answered with a full 'Use when the user says...' clause listing concrete trigger phrases. It also states an exclusion ('Not for polite brainstorming'), which sharpens the when. This matches the level-5 anchor exactly; level 4 would require the 'when' to be less explicit, which it is not.

Trigger Term Quality

Trigger coverage is comprehensive and natural: 'andreessen', 'pmarca mode', 'should I build this', 'is there a market', 'are we at product/market fit', 'pmf check', 'pressure-test this idea', 'be brutal about this venture', 'market-first take'. These are exact phrases a user would naturally say, including persona names, synonyms, and intent variants. Not level 4 because even uncommon variations ('pmf check', 'market-first take') are present, matching the comprehensive-with-synonyms anchor.

Distinctiveness Conflict Risk

The persona framing ('Marc Andreessen-mode', 'pmarca mode') creates a clear niche with minimal conflict, but a few trigger phrases are broad: 'should I build this' and 'is there a market' could plausibly match a generic startup-advice or decision-making skill, creating minor overlap risk with closely related skills. This fits the level-4 anchor ('mostly distinct; minor overlap risk') better than level 5, whose example shows triggers that essentially no other skill would claim.