agentleFS
Sign inStart free
tboome33tboome33/obsidian-mcp-router/skills/manage-delete/SKILL.md

manage-delete

Skill1 starsChanged 7 months agoApache-2.0View on GitHub

Permanently delete a file. Requires explicit confirm=true on a second invocation — first call shows a preview and refuses, second call with confirm=true proceeds. Designed against accidental deletes from hallucinated tool calls.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 0 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Call the obsidian-router delete_file MCP tool with arguments parsed from $ARGUMENTS.

    line 20
  2. Very specific:

    Ask: *"Confirm delete? Re-run with confirm=true to proceed."*

    line 42
  3. Very specific:

    DO NOT call delete_file with confirm=true on the user's behalf.

    line 43
  4. Very specific:

    If the user clearly types confirm=true from the start, you can call directly — but err on the side of confirming.

    line 49
  5. Very specific:

    If the call fails, do NOT delete the file with shell rm / Remove-Item on the vault's real path:

    line 55

≈ “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

manage-delete

Call the obsidian-router delete_file MCP tool with arguments parsed from $ARGUMENTS.

Arguments

Required:

  • path — file path relative to vault root.
  • confirm — must be EXACTLY true. Any other value blocks the deletion.

Optional:

  • vault — omit for default.
  • ifMatch — a contentSha256 from a prior get_file. Refuses the delete with a 409 conflict if the file changed since you read it. Use it when deleting a file you inspected earlier and another session could have edited in between — it stops you from deleting someone's just-made changes.

Argument parsing from $ARGUMENTS

  • bare path → path. DO NOT auto-set confirm=true. The first call should fail with the guard message; only set confirm=true when the user re-issues the command with explicit confirmation.
  • <path> --yes or <path> confirm=true → set confirm=true

Safety protocol

  1. First invocation without confirm=true:
    • Show the user the path that's about to be deleted
    • Optionally call get_file to show a preview (first 10 lines) so they know what they're deleting
    • Ask: "Confirm delete? Re-run with confirm=true to proceed."
    • DO NOT call delete_file with confirm=true on the user's behalf.
  2. Second invocation with confirm=true:
    • Call the tool
    • Report vault, path, deleted: true

If the user clearly types confirm=true from the start, you can call directly — but err on the side of confirming.

This guard exists because Claude can hallucinate delete calls in long sessions. The protocol gives the user one more chance to catch it.

On failure — remediate, NEVER fall back to filesystem operations

If the call fails, do NOT delete the file with shell rm / Remove-Item on the vault's real path:

  • Connection error (ECONNREFUSED, timeout) → vault closed; list_vaults, then ask the user to open it via the openUri link and wait.
  • Validation / API error (e.g. the confirm=true guard) → follow the safety protocol above; never bypass it via the filesystem.

Why FS deletes are worse here: they skip this skill's two-step confirm entirely. See the default-vault-health-check convention (canonical source).

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