← ClaudeAtlas

competitor-teardownlisted

Discovery-stage skill: turns material about a competitor's product — pricing pages, docs, changelogs, screenshots, customer reports — into a structured teardown where every claim is marked observed or inferred. Use when the user names a competitor product and asks for a teardown, competitive analysis, or comparison — 'tear down X', 'how does our Y compare to theirs', 'what is X doing with their pricing' — or when /pm routes such a request here. Do NOT use to write positioning or battlecards (Strategy stage), to scrape or collect the competitor data itself, for company-culture questions, or when no competitor material or product is identified — with nothing to observe it asks for material rather than inventing findings.
Abhillashjadhav/PM-agent-OS · ★ 1 · Data & Documents · score 60
Install: claude install-skill Abhillashjadhav/PM-agent-OS
# Competitor Teardown Material in, teardown out — with a bright line between what was seen and what is being guessed. ## Verification gates (defined first; output is blocked until all pass) - **G1 — Every claim tagged:** each factual claim carries exactly one of `[OBSERVED: <source>]` or `[INFERRED: <basis>]`. An untagged claim fails the gate. - **G2 — No invented facts:** every OBSERVED tag traces to material in the input (or material the skill actually fetched and quoted this session). Unknowns are written as "not determinable from input" — never filled with plausible numbers, features, or absences. - **G3 — Inferences grounded and unchained:** every INFERRED tag states its basis from observed material; an inference may not be the sole basis for a further inference presented as a finding. ## Steps 1. **Inventory the material.** List what was provided (pricing page, changelog, N customer reports, screenshots) — this inventory is the universe OBSERVED tags may cite. No material and no fetchable source? Stop and ask for material. 2. **Extract observations.** Pull concrete facts: prices, tiers, features named, release dates, quoted customer language. Single-source reports (one churned customer, one review) keep their provenance visible — "one customer reports X" is the observation; "X is true of the product" is not. 3. **Draw inferences — separately.** Cadence → investment level, pricing structure → target segment, missing tier features → upsell strategy. Each with its sta