agentleFS
Sign inStart free
Wsh7AshWsh7Ash/skillmcp/skills/destructive-action-confirmation/SKILL.md

destructive-action-confirmation

Skill0 starsChanged 15 days agoMITView on GitHub

Require a second, explicit signal before a destructive call leaves a trace that cannot be undone. Use when the server exposes a tool whose effect cannot be reversed by the caller

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

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Avoid when: every operation is reversible through the same interface

    line 16
  2. Distinctive:

    Test a mismatched confirmation and confirm the refusal

    line 25

≈ “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

Destructive Action Confirmation

Require a second, explicit signal before a destructive call leaves a trace that cannot be undone

Use when: the server exposes a tool whose effect cannot be reversed by the caller

Avoid when: every operation is reversible through the same interface

Procedure

  1. Classify each tool as reversible or irreversible
  2. For irreversible tools, require a confirmation argument derived from the target
  3. Refuse when the confirmation does not match the target exactly
  4. Return the effect that will occur before performing it when a dry run exists
  5. Log the confirmation value with the call
  6. Test a mismatched confirmation and confirm the refusal

Signals

  • an irreversible call without a matching confirmation is refused
  • the refusal states the expected confirmation value
  • the log records both the target and the confirmation
  • a dry run leaves no state change

Patterns

1. Confirmation token

confirm='delete pack:legacy' must equal the target

2. Refusal shape

confirmation mismatch: expected 'delete pack:legacy'

3. Dry run mode

same call with dry_run=true, no write

Pitfalls

  • Accepting a boolean confirm flag that any agent can set blindly
  • Confirming a category rather than the exact target
  • Skipping the dry run because the tool is internal

Done when

  • irreversible tools require a target-derived confirmation
  • a mismatched confirmation is proven to be refused

References

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