agentleFS
Sign inStart free
ndisisndndisisnd/wdym/SKILL.md

wdym

Skill1 starsChanged 4 months agoMITView on GitHub

Rewrite, improve, or enhance a prompt. Runs on /wdym, on any request to improve/enhance/rewrite a prompt, and on a UserPromptSubmit <prompt-detect> block — once per block (one prompt_id), never again for a Stop-hook bounce or retry in the same turn. Detects prompt type (code, question, text-gen), applies up to 3 matched principles, and runs the rewrite immediately — or runs the prompt as typed when nothing would improve it. Manage with /wdym --init, --status, --help.

2 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions2 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 2

    1. AmosChenZixuan/Agentic-working-contract★ 2same name
    2. jtlgrowth/wdym★ 1same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository3 other files it gives its agents.

The file

Create a free account to mount
Files
SKILL.md

wdym

Host portability. wdym runs on any agent host with a UserPromptSubmit hook. Two things vary by host while the skill is running (a third set — where installed files land — is resolved by refs/init.md, and only during --init):

  • Command prefix — /wdym on Claude Code, $wdym on Codex. This skill writes /wdym throughout; use the running host's prefix in anything shown to the user.
  • Asking the user (only during --init) — the ask step in refs/init.md. Use the AskUserQuestion tool when it is available, otherwise ask in plain text and end the turn. Choose by whether that tool is available, never by guessing which host you are on.

Activation is set by pref.json's activation key and chosen at /wdym --init:

  • hook — fires on UserPromptSubmit. <prompt-detect> block present ⇒ run this skill; absent ⇒ respond normally. Run it once per block: the block stays in context for the whole turn, so a Stop-hook bounce, a retry, or a later step of the same turn is not a new prompt. A block whose prompt_id already ran is done.
  • Plugin install — the plugin's hooks/hooks.json wires the hook, and an absent activation key means hook. /wdym --init still writes the pref and the contract, and never adds a settings entry (refs/init.md Step I5).
  • on-demand — no hook fires; prompts arrive with no block, which the protocol handles via its no-hook fallback paths.

Either way, an explicit /wdym <prompt> or a request to "improve / enhance / rewrite this prompt" runs the skill.

Follow refs/protocol.md end-to-end (Steps 0–7). If the prompt carries a /wdym command flag (--init / --help / --status, or a retired run-mode flag), follow refs/commands.md instead and terminate — never enhance a prompt. An inline --global flag is handled inside the protocol.

Output discipline: flash-only — no gate, no preview, no step markers, no original prompt, no principles or rationale. Visible output is only the self-check repair line, when something was healed.

Caching: refs/protocol.md holds the whole run, principles included — read it once per session and rebuild the working principle pool per run from it.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.

Check your setup, free