Agent Permission Boundary Audit
Automated governance and risk audit for AI agent tool permissions and authentication boundaries.
$5
Works with the AI Agents you already use
agent-permission-boundary-audit
Example session with this skill installed
Audit the Support Agent Pilot using sample-tools.json and sample-policy.md to check for unauthorized destructive permissions.
- Read your context and instructions
- Compiled the agent-permission-boundary-audit
Finding: Over-broad scope in 'SupportAgent'
Tool: 'db_delete'
Risk: High - No human-in-the-loop approval found in policy.md for destructive actions.
Hardening: Implement a mandatory approval gate for the 'delete' scope in connector-config.json.
Connects securely to your tools. The creator never sees your data.
About this skill
What it does
This skill provides a comprehensive security and governance audit for AI agent systems. It analyzes tool inventories, authentication models, connector scopes, and execution logs to identify over-privileged tools and risky permission combinations.
Why use this skill
Manual security reviews for LLM agents are prone to oversight, especially when tracking complex tool-calling boundaries. This skill automates the detection of "Shadow AI" risks and governance gaps by mapping your agent's actual capabilities against your defined security policies. It goes beyond simple prompting by cross-referencing multi-source evidence—including run logs and credential models—to ensure your safety guardrails are actually effective.
Supported Tools
The skill integrates with standard development environments using Python and PowerShell. It consumes JSON-based tool inventories, auth configurations, and policy documentation to generate machine-readable JSON audits and client-ready Markdown reports.
The Output
You receive a detailed privilege matrix and structured hardening plan. Findings are categorized by severity and tied directly to specific tool metadata or policy violations, providing a clear roadmap for securing your agent pilots.
- Does not perform live network penetration testing.
- Accuracy depends on the quality of provided JSON tool inventories and policy documentation.
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 goes through the trust checks below.
Security scanned
Checked by 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 scan
Security scan passed 7 months ago
Permissions
No special permissions declared or detected