marcindudekdev/consolidate-permission-prompts — explained in plain English
Analysis updated 2026-05-18
Clean up a Claude Code permissions.allow list that has grown to hundreds of rules.
Preview which permission rules are redundant or junk before removing anything.
Collapse a family of narrow allow rules into one wildcard rule when you choose to.
| marcindudekdev/consolidate-permission-prompts | 0xustaz/streamgate | a-bissell/unleash-lite | |
|---|---|---|---|
| Stars | 1 | 1 | 1 |
| Language | Python | Python | Python |
| Setup difficulty | easy | hard | hard |
| Complexity | 2/5 | 4/5 | 4/5 |
| Audience | developer | developer | researcher |
Figures from each repo's GitHub metadata at analysis time.
Requires Python 3.14 with no extra dependencies, just clone into the skills folder.
This is a skill for Claude Code, the AI coding assistant, that cleans up a list of permission rules that grows over time. Every time someone clicks allow always while using Claude Code, a new rule gets added to a permissions file. After a few hundred approvals, that list is full of clutter: narrow rules already covered by a broader one, one-off commands that will never run again, and stray fragments from shell loops that got captured by accident. This tool finds and removes that clutter. It works by reading your permission rules across three levels (user, project, and local, which all combine when Claude Code runs) and sorting each rule into one of four groups: redundant rules already covered by a broader rule, junk rules that will never match anything real, rules to keep as is, and optional consolidations where a family of narrow rules could be replaced by one wider rule. Removing redundant and junk rules does not change what is allowed, it just tidies the list. Consolidating rules is different because it can widen what is allowed, so that step is always a separate, explicit choice the user must confirm. Safety is built into how it works. The analysis script only reads and reports, it never edits your settings file on its own. It is careful about scope, so it will never delete a broader rule just because a narrower project-specific version matches, since that would quietly reduce what you are allowed to do. It also refuses to treat similar-looking commands as the same thing, so a rule for one command will not be mistaken for covering a different one. Any consolidation that would widen access to something risky, like running python, ssh, curl, or sudo, is specifically flagged as widening execution reach. The tool never touches deny or ask rules, only allow rules. To use it, you clone the folder into your Claude Code skills directory, then run the built-in slash command or the Python analysis script directly, with options to preview changes before applying them. It requires Python 3.14 with no extra dependencies and includes 96 tests. It is released under the MIT license.
A Claude Code skill that cleans up bloated permission-allow lists by removing redundant and junk rules.
Mainly Python. The stack also includes Python.
Use, modify, and share freely for any purpose, including commercial use, as long as you keep the copyright notice.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Don't trust strangers blindly. Verify against the repo.