{"slug": "i-made-my-coding-agent-audit-my-prompt-before-it-writes-anything", "title": "I made my coding agent audit my prompt before it writes anything", "summary": "A developer built an open-source agent skill called `clarify` that audits a coding prompt for silent assumptions before the agent begins work, surfacing only the gaps that matter and pausing until the user answers. The skill, available free on GitHub and usable as a standing instruction file or a `/clarify` command, sorts each assumption into three bands so that only open questions become prompts. The developer argues that the real problem is not ambiguity but silent inference, since a compressed expert prompt reflects unstated intent rather than absent intent.", "body_md": "The cost of a vague prompt is paid at the end, not the start. You learn what the model assumed only after it has finished building on the assumption.\n\nThat is the problem worth solving. You write down your intent, the agent acts on what you wrote, and whatever you left out gets inferred. The gaps never announce themselves. The model reaches for the reading its training makes likeliest, which is a decent guess about the average requester and a worse guess about you.\n\nSo I built a skill called `clarify`. It is a set of standing instructions the agent reads first, and it can also be called as a `/clarify` command. Before work starts, it audits the prompt for every assumption it would otherwise fill silently, surfaces them, and does not proceed until I answer.\n\nAuditing every gap would be unbearable. A prompt has dozens and most of them do not matter. So the skill sorts each assumption into three bands:\n\nOnly the open ones become questions. That is the difference between a clarifying step you keep and one you switch off after a day.\n\nFrom the skill file:\n\nThe enemy is not ambiguity, it is silent inference. An expert's prompt is compressed: what they don't say is not absence of intent, it is unstated intent. Recover it from them; do not reconstruct it from your priors.\n\nA compressed prompt is not a bad prompt. Treating compression as absence is what produces confidently wrong work.\n\nPublic, free, no signup: [https://github.com/bryanthood-wph/clarify](https://github.com/bryanthood-wph/clarify)\n\nDrop it in and call it. It works as a skill file, and the `/clarify` command is the same instructions on demand.\n\nThe longer argument, on why a prompt is the cleanest record of your own blind spots, is over here: [https://medium.com/@brnthood/account-for-what-your-prompt-doesnt-say-b443fa3af231?sk=504afa509c9d9df58dbff3290df992a3](https://medium.com/@brnthood/account-for-what-your-prompt-doesnt-say-b443fa3af231?sk=504afa509c9d9df58dbff3290df992a3)", "url": "https://wpnews.pro/news/i-made-my-coding-agent-audit-my-prompt-before-it-writes-anything", "canonical_source": "https://dev.to/bryant_hood_5d0e171a08fec/i-made-my-coding-agent-audit-my-prompt-before-it-writes-anything-56c5", "published_at": "2026-09-21 13:14:15+00:00", "updated_at": "2026-09-21 13:27:13.400128+00:00", "lang": "en", "topics": ["ai-agents", "ai-tools", "developer-tools", "large-language-models"], "entities": ["GitHub", "Medium", "clarify"], "alternates": {"html": "https://wpnews.pro/news/i-made-my-coding-agent-audit-my-prompt-before-it-writes-anything", "markdown": "https://wpnews.pro/news/i-made-my-coding-agent-audit-my-prompt-before-it-writes-anything.md", "text": "https://wpnews.pro/news/i-made-my-coding-agent-audit-my-prompt-before-it-writes-anything.txt", "jsonld": "https://wpnews.pro/news/i-made-my-coding-agent-audit-my-prompt-before-it-writes-anything.jsonld"}}