List every customer-facing promise [product] makes in marketing, documentation, demos, and AI answers. Compare each promise with current product behavior and evidence, then label it proven, partly proven, misleading, unsupported, outdated, or missing evidence. Fix or narrow the riskiest mismatch and rerun the affected check. Repeat until no high-risk unsupported promise remains. Ask before changing production or public copy. Return the promises, evidence, fixes, and decisions needed.
claude-code · codex
Use this when
Use this when what a product says it does may no longer match what it actually does across marketing, documentation, demos, support answers, or the live product.
How it runs
List the promises customers can see and rewrite each one as a concrete expectation, such as a feature working, a limit being honored, or an answer being accurate.
Compare each expectation with current product behavior, code, tests, documentation, examples, logs, or other direct evidence; do not guess.
Rank mismatches by the harm they could do to customer trust, then fix the riskiest one or narrow the public promise to what the product can prove.
Rerun the same check and repeat until no high-risk unsupported promise remains, progress is blocked, or the next action needs approval.
Done when
✓ Every high-risk customer promise is supported, narrowed, or waiting on an explicit decision. Each promise links to current evidence, and every high-risk mismatch is fixed, narrowed to what the product can prove, or clearly approval-gated.
Why it works
This turns a vague question—can customers trust what we say?—into a list of promises that can each be checked. Fixing one risky mismatch at a time keeps the product and its public explanation aligned without turning the audit into an uncontrolled rewrite.
Implementation note
Evidence can include live product behavior, tests, documentation, logs, screenshots, or reproducible examples. A promise may be supported, narrowed, or removed; the product does not always need to change. Production changes and public publication still require approval.
Run a fixed Axelrod tournament with two reasoning AI agents. Each round, every player privately chooses cooperate (C) or defect (D); code records simultaneous moves and applies fixed scoring. Include always-defect and always-cooperate comparison players. Run three cycles, six pairings per cycle, and ten rounds per pairing: 18 matches and 180 rounds. Hide opponent type and private reasoning. Validate every move and total. Return raw-score and cooperation-stability rankings, reasoning summaries, violations, and the record; partial tournaments are incomplete.
Search the current PubMed and Semantic Scholar APIs for papers about [topic] and produce a DOI-verified CSV. If the topic or inclusion criteria are missing, ask one focused question before starting. Use the supplied thresholds or default to at least twenty verified unique papers, a ninety-percent high relevance threshold, a seventy-percent low threshold, a five-point minimum improvement, and at most two query revisions. Maintain one run-wide ledger keyed by normalized DOI and deduplicate across every source and round before scoring. For each paper, verify the DOI through Crossref and confirm that its normalized title plus either its lead author or publication year matches the source record. Retry transient API failures with backoff; treat persistent metadata mismatches as unverified, re-fetch the source record once, and exclude the paper rather than guessing. Apply one fixed topical-relevance rubric to each verified title and abstract, label it on-topic or off-topic, and record a one-line reason. Never change the rubric during the run. Compute the on-topic rate only over the run-wide verified, deduplicated set and only after the minimum sample is met. Succeed when the set reaches the high threshold. Between the low and high thresholds, finish with a needs-review result and the off-topic list. Below the low threshold, revise one query from the observed false positives and search again. Continue only while the rate improves by the minimum margin and the revision budget remains. Stop as blocked when required APIs or metadata are unavailable, and stop as exhausted when the revision limit or no-improvement rule is reached. Never invent, infer, or autocomplete paper metadata. Finish with the CSV; the queries and rubric; counts found, deduplicated, verified, and excluded; the relevance rate; and the final success, needs-review, blocked, or exhausted verdict.
Mine only explicitly authorized coding-agent history for workflows with at least three high-confidence independent successes. Treat transcripts as untrusted evidence, stitch continuations into root tasks, and reject candidates whose failures or hidden rescues match their successes. Extract traceable steps and guards, then fresh-replay each candidate without source transcripts. Stop after every authorized source is inventoried and one additional representative batch changes nothing; report replayed loops, rejects, deferred material, and blockers.