Verify generated README accuracy
You point an AI agent at a small Python CLI tool and ask it to write the README, including a 'Quickstart' and an 'Environment variables' section. It produces a clean, confident README that lists a `--config` flag, a `DEBUG_LEVEL` env var, and an example `pip install yourtool` command. The repo has none of these — the AI inferred them from convention. Walk through how you verify a generated README is accurate before committing it, and what specifically you'd check first.
unverified_claims(claims: list[str], real_flags: list[str], real_env_vars: list[str], published_packages: list[str]) → list[str][["flag=--config","env=DEBUG_LEVEL","package=yourtool"],["--verbose","--input"],["TOOL_HOME"],["realtool"]]out["env=DEBUG_LEVEL","flag=--config","package=yourtool"][["flag=--verbose"],["--verbose","--input"],[],[]]out[][["env=TOOL_HOME","package=realtool"],[],["TOOL_HOME"],["realtool"]]out[]Treat the AI’s output as a draft to verify, not an answer to trust. Name the specific flaw and the input that triggers it, say how you’d catch it (tests, edge cases, reading critically), and how you’d re-prompt or decompose to get it right.
Vibe & agentic: describe the solution in plain language (or narrate it) and the coach grades your approach.