Andreahlert Scope Guard
andreahlert-scope-guard is a Claude Code plugin for the Ship phase that detects when your AI coding agent exceeds the scope you asked for and applies guardrails via hooks.
by andreahlert · github.com/andreahlert/scope-guard
Install scope-guard so Claude Code hooks detect and block agent changes that exceed what you actually requested.
Add it to Claude Code
Install the plugin in Claude Code. One command, paste-ready.
/plugin install andreahlert-scope-guard@andreahlert/scope-guardBuilt to be called by your agent
Skillselion is itself an MCP server. Your agent can pull this entry and a paste-ready install config straight from the API - no copy-paste.
Retrieve this entry with skillselion.get_details("plugin:andreahlert/scope-guard") and the paste-ready config with skillselion.get_install_config("plugin:andreahlert/scope-guard").
What it does
andreahlert-scope-guard is a Claude Code plugin that watches your agent for scope creep—the moment the model edits files, adds dependencies, or expands tasks beyond what you asked. Solo builders who delegate multi-file refactors know how quickly a small bugfix becomes a framework migration; this plugin is positioned as guardrails plus hooks that detect that overreach early. The catalog lists developer-tools keywords including agent, coding, detects, guard, guardrails, hooks, and scope-creep, which signals runtime enforcement rather than a passive checklist skill. Journey-wide placement fits because you can invoke the same protection during validate prototypes, build integrations, and pre-release review, but the primary shelf remains ship/review where unapproved diffs are most costly. Install it when you trust the agent for speed but not for boundary discipline. It complements planning skills; it does not replace writing a spec—it gives you automated pushback when execution diverges from the brief.
Highlights
- Detects when the coding agent goes beyond the user’s stated request
- Developer-tools plugin with hooks and guardrails for scope creep
- Surfaces scope, dependencies, and config drift during agent sessions
- Targets AI scope creep called out explicitly in the marketplace description
- Single-plugin bundle from andreahlert/scope-guard repository
Why builders use it
Indie developers waste review time reverting drive-by refactors, mystery dependencies, and unrelated files the agent touched while “fixing” one request.
After registration, scope-guard can interrupt or flag out-of-scope agent actions so you approve only the work you actually commissioned.
At a glance
- Type - Plugin in Development Tools.
- Adoption - 0 installs, 2 stars, 0 votes.
FAQ
Who is andreahlert-scope-guard for?
It is for Claude Code users who delegate coding tasks and want automated detection when the agent wanders past the original request.
When should I use andreahlert-scope-guard?
Use it whenever the agent can touch multiple files or dependencies—especially before you merge or ship changes you did not explicitly ask for.
How do I add andreahlert-scope-guard to my agent?
Add the andreahlert/scope-guard plugin from the marketplace in Claude Code, configure hooks and guard settings per repo, then run agent tasks under those guardrails.
Comments
Share how you use andreahlert-scope-guard, gotchas, or tips for other indie builders.
No comments yet - be the first to share how you use it.