Agent skill

vibe-scope-guard

Detects and redirects scope creep during implementation. Catches unrequested features, unnecessary refactoring, and over-engineering.

Stars 163
Forks 31

Install this agent skill to your Project

npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/other/other/scope-guard-ash1794-vibe-engineering

SKILL.md

vibe-scope-guard

The best code is the code you don't write. Stay focused on what was asked.

When to Use This Skill

  • During any implementation task
  • When you notice you're "improving" code that wasn't in the request
  • When implementation is taking longer than expected
  • When you're about to create an abstraction "for later"

When NOT to Use This Skill

  • During brainstorming (ideas should be unconstrained)
  • When the user explicitly asks for broad improvements
  • When scope expansion is necessary for correctness

Scope Creep Signals

Watch for these patterns:

Signal Example Response
Unrequested features "While I'm here, let me add caching" Stop. Was caching requested?
Premature abstraction "Let me create a generic helper for this" Is it used more than once?
Gold plating "Let me add comprehensive error messages for every case" Only at system boundaries
Refactoring drive-by "This function could be cleaner" Was refactoring requested?
Over-engineering "Let me make this configurable" Does anyone need configuration?
Documentation creep "Let me add docstrings to all these functions" Only to functions you changed
Test over-expansion "Let me test every possible input" Test the boundaries, not every input

The Rule

Three similar lines of code is better than a premature abstraction.

Before adding anything not explicitly requested, ask:

  1. Was this requested? → No → Don't do it
  2. Is it necessary for correctness? → No → Don't do it
  3. Will it break without this? → No → Don't do it
  4. Is the user watching time/cost? → Yes → Definitely don't do it

Steps

When you detect scope creep:

  1. Stop — Don't commit the extra work
  2. Acknowledge — "I notice I'm adding [X] which wasn't requested"
  3. Offer — "Would you like me to also [X], or should I stay focused on [original task]?"
  4. If declined — Revert the extra work, continue with original scope
  5. If accepted — Proceed, but track it separately

Output

This skill doesn't produce formal output. It modifies behavior. When triggered explicitly:

Scope Audit

Original request: [what was asked] Current work: [what's being done] Scope additions detected: X

Addition Requested? Necessary? Recommendation
[extra work] No No Remove

Expand your agent's capabilities with these related and highly-rated skills.

Didn't find tool you were looking for?

Be as detailed as possible for better results