Works with the AI Agents you already use
Claude Code
Cursor
Codex CLI
Muse
Dots+22 moreClaude CodeCursorCodex CLIMuseDots
OpenClaw
Manus
Grok
Grok Bot+18 agentsTool Permission Scope Auditor
Audit delegated tool permissions to ensure least-privilege access and secure workflow execution.
$12.99
Tool Permission Scope Auditor
Example session with this skill installed
Audit a GitHub Action workflow that needs to read issues and post comments. Current credential is a Personal Access Token with 'repo' scope. Environment is production.
- Read your context and instructions
- Compiled the tool permission scope
The audit for the GitHub Action workflow is complete. The current configuration is heavily over-scoped.
Permission Scope Control Pack
Executive Decision: NOT_READY
- Objective: Read issues and post comments on production repositories.
- Verdict: CRITICAL Over-scope detected.
Blocker: Personal Access Token (PAT) with full repo scope provides destructive and administrative authority not required for this task.
1. Permission Scope Evidence & Least-Privilege Ledger (PSELL)
| PSELL ID | Tool | Intended Action | Consequence Class | Evidence State | Least-Privilege Decision | Required Remediation |
|---|---|---|---|---|---|---|
| PSELL-01 | GitHub API | read_issues | READ_ONLY | CONFIRMED | OVER_SCOPED | Use GitHub App with metadata:read |
| PSELL-02 | GitHub API | create_comment | REVERSIBLE_WRITE | CONFIRMED | OVER_SCOPED | Use GitHub App with issues:write |
| PSELL-03 | GitHub API | delete_repo | DESTRUCTIVE | CONFIRMED | PROHIBITED | Revoke PAT; PAT has full repo control |
2. Tool-Scope Matrix
- Tool: GitHub REST API
- Principal: User PAT (repo scope)
- Allowed Read: All repository content, code, and metadata.
- Allowed Write: Commits, branches, and releases.
- Allowed Delete: Can delete the entire repository.
- Remediation Priority: CRITICAL.
3. Remediation & Recovery Plan
- Reduce Scope: Replace the PAT with a GitHub App installation token.
- Technical Restriction: Grant
issues:readandissues:writepermissions only.
Resource Boundary: Restrict the App installation to the specific production repository instead of all user repos.
4. Validation: Attempt to delete a branch using the new token; the request must return a 403 Forbidden status.
Next steps
- Generate a GitHub App with granular permissions.
- Replace the
GH_TOKENsecret in the repository settings. - Run the validation test to confirm the 403 response on destructive actions.
Connects securely to your tools. The creator never sees your data.
About this skill
Technical debt and excessive permissions often lead to catastrophic security breaches. This skill converts raw permission data and workflow requirements into a traceable, implementation-ready control pack. It moves beyond simple "access granted" flags to analyze effective permissions, inherited roles, and human approval boundaries.
What it does
- Permission Auditing evaluates technical capabilities against effective access to find gaps or excessive authority.
- Least-Privilege Mapping derives the minimum required permissions for specific delegated tasks.
- Evidence Verification classifies access claims by strength, distinguishing between confirmed facts and inferred guesses.
- Remediation Planning prescribes exact technical steps to reduce over-scope or fix under-scoped accounts.
- Approval Gating separates technical access from delegated authority, ensuring human-in-the-loop controls for high-risk actions.
How it works
- Supply the task objective, available tools, current credentials, and any available evidence like role exports or API scopes.
- The agent generates a Permission Scope Evidence & Least-Privilege Ledger (PSELL) to identify every risk and blocker.
- Review the remediation plan and readiness checklist to see which parts of the workflow are safe to execute.
Frameworks & tools
Works with any identity provider or platform including AWS IAM, Azure RBAC, GCP Service Accounts, GitHub Actions, Slack App scopes, and custom OAuth2 implementations.
Why this beats prompting it yourself
Generic prompts often miss inherited permissions or hidden platform behaviors that increase your blast radius. This skill enforces a strict Evidence Strength Hierarchy and a multi-dimensional core control model that treats technical permission and delegated authority as separate requirements.
Use cases
- Auditing a new AI agent's service account before deployment to production.
- Validating that a CI/CD pipeline has read-only access to specific buckets rather than admin rights.
- Reviewing temporary access exceptions for contractors or external services.
Known limitations
Does not execute permission changes directly unless an environment is provided. It cannot guess permissions for hidden or undocumented proprietary systems.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill passes our trust checks below.
Security scanned
Passed our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scanned
Verified clean 19 days ago
- Passed all security checks, Safe to install