Scope Creep Detector
Flags what your agent added that you didn't ask for.
What it does
It compares a plan or a Git diff with your original request and lists anything that goes beyond it, such as new features, refactors you didn't ask for, extra dependencies, or abstractions for needs you don't have yet.
Why it's useful
Ask a coding agent for a CSV export and you can end up with an export framework that also does PDF, a new package, and a rewritten table component. Any one of those might be defensible, but together they make a much bigger diff to review and more code to maintain.
The skill goes through each change and tries to tie it back to something you asked for. When it can't, it flags the change, explains why, and suggests what to do about it. Whether it stays is up to you.
Tests, security and accessibility fixes count as justified even if you didn't mention them, so it doesn't treat those as extras.
How it sorts each change
- Required
- Directly implements something you asked for.
- Justified
- Not asked for, but needed to ship it safely. The report says why.
- Questionable
- Might be useful, but nothing in the request backs it up yet.
- Out of scope
- Has nothing to do with the request and can usually be dropped.
Audit mode (default)
It only reads and reports. It doesn't edit or revert anything, so what happens next is up to you.
Strict mode (ask for it)
The agent stops and asks you before doing anything the skill would flag. This is an instruction to the agent rather than something enforced, so it depends on the agent following it.
What a report looks like
illustrative example
Scope baseline: Add a CSV export button to the existing users table.
Verdict: Some expansion
Findings
Out of scope · src/components/UsersTable.tsx
Column sorting and layout were rewritten. Export doesn't need it.
Fix: restore the original table, keep only the new button.
Questionable · src/lib/export/
Generic exporter with PDF and XLSX adapters. Only CSV was asked for.
Fix: a small CSV helper. Add formats when someone needs them.
Keep: CSV escaping, regression test for exported rows.
Next action: drop the exporter abstraction, then re-run the audit.Install
The skill is a folder with a SKILL.md in it, following the Agent Skills format. There's nothing to build or run, and it doesn't need an API key.
01Claude Code
Clone it into your personal skills folder to use it in every project.
$ git clone https://github.com/ritenv/scope-creep-detector ~/.claude/skills/scope-creep-detectorTo use it in a single repo, clone it into
.claude/skills/scope-creep-detectorat the project root instead. If you commit that folder, everyone working on the repo gets the skill.02Claude apps
Zip the folder and upload it as a custom skill from Settings in claude.ai or the desktop app.
$ git clone https://github.com/ritenv/scope-creep-detector && zip -r scope-creep-detector.zip scope-creep-detector -x "*.git*"03Codex, Cursor and other agents
If your agent supports Agent Skills, drop the folder into its skills directory. If it doesn't, paste the contents of
SKILL.mdinto your project instructions (AGENTS.md, a rules file, or a system prompt) and the agent can follow it from there.
What's in the folder
- scope-creep-detector/
- SKILL.md
- README.md
- references/classification-guide.md
Example prompts
Agents can pick the skill up on their own when a task matches its description, but naming it in your prompt is more reliable.
- Use scope-creep-detector to audit this plan against my original request.
- Audit my changes against the original request in strict mode.