what-huh
Skill1 starsChanged 9 days agoMITView on GitHub
Make final responses understandable to a reader who cannot see the agent's working text, especially when reporting progress, explaining findings, or asking for a decision.
In this repository23 other files it gives its agents.
- AGENTS.mdAGENTS.md
- aftermath-rewindSkill
- aiconographerSkill
- ask-socratesSkill
- check-project-nameSkill
- goal-checkSkill
- lean-proof-reviewSkill
- parallax-imageSkill
- prior-artSkill
- projectbotSkill
- quiet-uxSkill
- rmddSkill
- rmdsSkill
- rmdvSkill
- rrr-review-remedy-respondSkill
- soft-skill-creatorSkill
- thought-rigSkill
- todoSkill
- verifier-systemSkill
- what-whatSkill
- write-humanlySkill
- write-plainlySkill
- write-simplified-technical-englishSkill
The file
Files
What Huh
The reader sees your final response, not your thinking, tool output, or working commentary. Make the response stand on its own and move the work forward. Imagine a capable person arriving at the end of the work and asking, “What happened, what does that mean for me, and what needs my attention?”
Lead with the outcome or answer. Give the context needed to understand it: name the thing you are discussing, what changed or was found, and why it matters. When reporting a problem or uncertainty, explain its practical consequence. When a decision is needed, explain what the choice would change and give enough evidence to choose; make your recommendation clear when you have one.
Before sending, notice references that depend on invisible work: “this,” “the issue,” “the other option,” “as mentioned above,” or a bare filename, error, or status. Replace or anchor them with the relevant name and a little context. Links can support an explanation, but do not make the reader open one just to understand your point.
Be concise without making the reader reconstruct the story. Include the facts that affect the outcome or decision, including a material limitation or unfinished step. Leave out the running log and details that do not help. Match the response to the task: a simple answer may need only a sentence; a complicated result may need a few connected paragraphs or a short list.
For example, instead of “Fixed it. The second check still fails,” say “The login page now accepts valid passwords. The password-reset test still fails because the email service is unavailable, so we have not verified reset emails.” Use only facts actually established by the work; do not invent an explanation to make the story flow.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

