agentsclimarketplace
Sign in

Respect the guard

Skill jcdavis131/cursor-agent-skills/skills/respect-the-guard

When blocked by a safety guard (push-to-main protection, destructive-command confirm, permission boundary, failed eval gate), uses the legitimate path instead of bypassing — and surfaces exactly how the operator relaxes the guard if desired. Use when a guard stops the fast path, when tempted to use --no-verify or --force, when a deploy script offers only an exec variant, or when input/API limits apply.From its SKILL.md

Install
npx -y skills add jcdavis131/cursor-agent-skills --skill respect-the-guard

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.2 KB, 449 tokens by cl100k_base, as published. Nobody here has run it

Respect The Guard

A guard blocking the fast path is doing its job. Bypassing it is faster and wrong. Use the legitimate path, finish the work, and tell the operator how to relax the guard if they want to.

The move

  1. Don't bypass. No --no-verify, no --force to main, no manual confirmation override, no --amend to skip a hook, no renaming a command to dodge a pattern match.
  2. Use the legitimate path.
    • Push-to-main blocked → branch + PR (logical-commit-split).
    • Destructive command needs confirm → ask, then proceed (shell-confirm-hygiene).
    • Tool missing (e.g. gh) → produce the manual alternative (compare URL) in one step.
    • Permission denied → surface it; don't retry with a workaround.
  3. Surface the relaxation; don't take it. Tell the operator exactly how to relax the guard; they ratify, you never apply it unilaterally.
  4. Finish via the legitimate path. Branch pushed, PR URL produced, manual step is one click — work done; bypass skipped.

When to ask the operator to relax a guard

  • Guard is over-broad for this repo (solo owner, PRs are theater).
  • Legitimate path is genuinely impractical.
  • Propose relaxation, get yes, document it — not silent.

Anti-patterns

  • --no-verify to skip a failing hook — fix the issue.
  • --force to main — one-way door.
  • Renaming a command to dodge a deny pattern.
  • Applying a relaxation silently.
  • Failing the task because of the guard — legitimate path completes the task.

Extended patterns

Credential scope, stored-credential reuse, copy guardrails, API input limits, :exec script pairs, non-overridable gates: reference.md

Pair with

  • shell-confirm-hygiene · logical-commit-split · pre-commit-hygiene · readiness-report · validate-gate

What ships with it: 1 file

2.1 KB alongside SKILL.md

Keep looking

Skills are one crate of 326,834. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.