agentleFS
Sign inStart free
linxulelinxule/carrel/skills/env-doctor/SKILL.md

env-doctor

Skill2 starsChanged 7 days agoMITView on GitHub

Guided recovery for environment.json drift. Use when the researcher runs /carrel-fix, mentions environment drift, broken setup state, or wants help repairing conflicting settings.

7 other projects use 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. 0 are common, 6 are distinctive, and 4 are very specific to this project.

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 6

    Distinctive

    This author's own choices. Worth a look.

  • 4

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Always work in order:

    line 12
  2. Very specific:

    If status is valid, say so plainly and stop.

    line 16
  3. Very specific:

    Only run carrel env fix --safe --vault .

    line 21
  4. Distinctive:

    If validation or the dry-run reports deferred items, shift into guided recovery.

    line 24
  5. Distinctive:

    Use plain language. Avoid "ValidationError", "enum", or other implementation terms unless the researcher asks.

    line 29
  6. Distinctive:

    If the ambiguity is about sensitivity or cloud consent, explain the practical consequence: "This setting decides whether cloud tools are …

    line 31
  7. Very specific:

    If CLAUDE.md and environment.json disagree, say explicitly that environment.json is the structured source of truth, but ask which one …

    line 33
  8. Distinctive:

    Use short, concrete questions.

    line 38
  9. Very specific:

    After every applied change, append an entry to _meta/capability-log.md that includes:

    line 46
  10. Distinctive:

    Use a compact format, for example:

    line 53

“Common” is measured: kept by at least 6 in 10 of the 5 edited versions of this file we could read. “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

env-doctor

Diagnoses .carrel/environment.json, offers safe repairs with approval, and walks the researcher through ambiguous drift without hiding the tradeoffs.

Mode Detection

Always work in order:

  1. Layer 1: Validate Run carrel env validate --vault . --format json. If status is valid, say so plainly and stop.
  2. Layer 2: Safe fix preview Run carrel env fix --safe --dry-run --vault . --format json. Group the safe fixes into a short plain-language summary and ask whether to apply them. Only run carrel env fix --safe --vault . after explicit approval.
  3. Layer 3: Interactive doctoring If validation or the dry-run reports deferred items, shift into guided recovery. Explain the ambiguity in ordinary language, present 2-3 options, and ask the researcher to choose.

Interaction Rules

  • Use plain language. Avoid "ValidationError", "enum", or other implementation terms unless the researcher asks.
  • Group related fixes. Do not dump every field unless the researcher wants the raw detail.
  • If the ambiguity is about sensitivity or cloud consent, explain the practical consequence: "This setting decides whether cloud tools are okay by default."
  • If CLAUDE.md and environment.json disagree, say explicitly that environment.json is the structured source of truth, but ask which one reflects the researcher's current intent before changing either.
  • Re-run carrel env validate --vault . --format json after any applied fix to confirm the vault is clean.

Questions for Ambiguous Drift

Use short, concrete questions. Examples:

  • "Your sensitivity setting uses an older label. Should I treat this as medium sensitivity, which keeps cloud tools off by default?"
  • "CLAUDE.md says cloud tools are okay, but environment.json says no. Which one is current?"
  • "This tool is marked configured in your profile, but it is not installed on this machine. Should I mark it unavailable here?"

Logging

After every applied change, append an entry to _meta/capability-log.md that includes:

  • Timestamp
  • What changed
  • Why it changed
  • How to revert it

Use a compact format, for example:

- 2026-04-20 14:03 UTC — Updated `.carrel/environment.json` sensitivity from `cautious` to `medium`; revert by restoring `.carrel/environment.json.bak`.

References

  • references/fix-catalog.md for the safe-fix rules supported by Layer 2
  • /carrel-fix for the slash-command wrapper

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