Benchmark Due Diligence

What it does

Instead of a neutral briefing, this skill assumes the benchmark's story is inflated until proven otherwise, then runs a four-phase orchestration:

• Phase 0 — foundation by evidence: never infer ownership from shared names or domains; verify the real entity graph and, above all, the attribution of the headline stat ("took product X from 0 → 1M users" — founder, or the growth lead who left?). It also separates what you actually own from what merely belongs to your clients.
• Phase 1+2 — collect → verify pipeline: parallel collection agents attach a source URL and source_kind (self-reported/marketing vs independent third party) to every finding; adversarial verification agents grade evidence L1–L4 and issue verdicts (confirmed / mostly credible / doubtful / debunked-bubble), actively hunting falsifying evidence.
• Phase 3 — attribution breakdown: what share of the success is product vs market timing vs personal-IP marketing vs operations, with explicit separation of replicable method from luck and non-transferable endowment. Includes a bubble-busting table sorted most-water-first.
• Phase 4 — what this means for YOU: a resource-mapping table tagging each playbook element ✅ borrow-able / ⚠️ not replicable / 🔄 already doing / 🚫 bubble-don't-copy, plus landing points and an action list.

A notable design detail: a two-channel injection rule keeps your private context (client names, strategic intent) out of any agent that performs web searches — it reaches only the final mapping phase.

Who it's for

• Founders who suspect a rival's traction numbers are mostly PR
• Investors and strategy leads who need headline claims (Product Hunt #1, user counts, valuations) stress-tested
• Marketers who want to copy a creator's growth formula but need to know which parts are luck
• Consultants who want benchmark reports to end in an action plan, not a summary

Examples

1. "This founder claims 0→1M users — is that real?" The skill makes headline-claim attribution the top verification target and cross-checks against third-party sources.
2. Teardown of a competitor's distribution playbook: platform matrix, persona, content types, how they borrow other people's audiences, and how personal IP funnels into the product.
3. Pre-investment screen: entity and funding checks (with qcc for mainland-China registration data), a structurally similar peer comparison, and how this class of playbook typically fails.

Install the skill I found on Claude Skill Mart.
Copy the benchmark-due-diligence folder from the GitHub repo daymade/claude-code-skills into my ~/.claude/skills/benchmark-due-diligence/.
When it's done, tell me in one line what this skill can do.