agentleFS
bex-cobex-co/bex/.claude/skills/dont-code-but-answer/SKILL.md

dont-code-but-answer

Skill578 starsChanged 6 days agoApache-2.0View on GitHub

Answer-only mode. Investigate the codebase as needed to answer the user's question, but make no code or file changes. Use only when the user explicitly invokes /dont-code-but-answer or $dont-code-but-answer.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 9 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈9

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Do not create, edit, move, or delete any file: source, config, tests, docs, generated output, or scratch files inside the repo.

    line 13
  2. Very specific:

    Do not run commands that change state: installs, builds, codegen, formatters, linters with --fix, migrations, snapshot updates, or git …

    line 14
  3. Distinctive:

    Do not change remote state: deploys, API writes, PRs, issues, or comments.

    line 15
  4. Distinctive:

    Do not hand the work to a subagent or background job that would do any of the above.

    line 16
  5. Distinctive:

    Read files, search the codebase, and inspect history (git log, git show, git blame, git diff).

    line 21
  6. Distinctive:

    Run read-only commands that observe behavior without writing to the working tree.

    line 22
  7. Distinctive:

    Look up external documentation when the question needs it.

    line 23
  8. Distinctive:

    If the answer implies a change, say what you would change and where: the file, the function, and the gist of the edit.

    line 31
  9. Distinctive:

    When there are real tradeoffs, recommend one option instead of surveying all of them.

    line 32
  10. Distinctive:

    Stop after answering. Don't start implementing.

    line 34

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Don't code — answer

The user wants an answer, not a change. Treat the text sent with this invocation as the question; if nothing was sent, answer the user's most recent question. Stay in this mode for follow-up questions until the user explicitly asks you to make a change.

Hard limits

  • Do not create, edit, move, or delete any file: source, config, tests, docs, generated output, or scratch files inside the repo.
  • Do not run commands that change state: installs, builds, codegen, formatters, linters with --fix, migrations, snapshot updates, or git commands that commit, check out, stash, reset, rebase, or push.
  • Do not change remote state: deploys, API writes, PRs, issues, or comments.
  • Do not hand the work to a subagent or background job that would do any of the above.
  • This holds even when the fix is a one-liner. Describe it; don't make it.

Allowed

  • Read files, search the codebase, and inspect history (git log, git show, git blame, git diff).
  • Run read-only commands that observe behavior without writing to the working tree.
  • Look up external documentation when the question needs it.

Investigate enough to answer correctly. Don't guess about code you can read.

How to answer

  • Lead with the direct answer in the first sentence or two, then the evidence.
  • Ground non-obvious claims in the code with path/to/file.ts:42 references.
  • If the answer implies a change, say what you would change and where: the file, the function, and the gist of the edit. A short snippet is fine when it is clearer than prose; keep it illustrative, not a full implementation.
  • When there are real tradeoffs, recommend one option instead of surveying all of them.
  • Say plainly what you could not verify.
  • Stop after answering. Don't start implementing; the user will ask for the change if they want 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.