- Home
- Skills
- Legal & Compliance
- Permission Aware Knowledge Finder
Works with the AI Agents you already use
Permission Aware Knowledge Finder
Enforce role-based access control and automatic redaction for sensitive internal company documents.
$5
Permission Aware Knowledge Finder
Example session with this skill installed
User: jdoe (Role: Sales Associate). Query: What is the current salary for the Senior Account Manager role and what are the performance bonuses for the Sales team?
- Read your context and instructions
- Compiled the permission aware knowledge
I have processed your request regarding compensation and bonuses, filtering the results based on your Sales Associate permissions.
[Permission-Aware Response]
🔒 Access Check:
- User Role: Sales Associate
- Required Clearance: HR / Management
- Status: PARTIALLY AUTHORIZED
📄 Answer:
Regarding the Sales team performance bonuses, all members of the Sales department are eligible for a quarterly bonus based on meeting 100% of their individual quota. The current bonus structure is calculated at 10% of the base salary for that period, paid out in the following month's payroll cycle.
Regarding the specific salary for the Senior Account Manager role: I cannot fulfill this request. The information you are asking for requires HR Admin clearance, and your current role is Sales Associate. Please contact your manager or the HR Department for assistance.
[Self-Check]
- User role explicitly checked against data sensitivity
- No external knowledge or guessing used for sensitive topics
- Redaction protocol applied correctly if partially authorized
- Hard refusal template used if fully denied
Next steps
- Review the Sales Handbook for detailed quota tiers.
- Contact HR for a full breakdown of management-level compensation bands.
Connects securely to your tools. The creator never sees your data.
About this skill
The problem
Sharing company knowledge via AI often leads to accidental data leaks. Standard RAG systems often ignore user permissions, surfacing sensitive HR, financial, or legal data to unauthorized employees.
What it does
- Enforces role-based access control (RBAC) on all AI-retrieved information.
- Cross-references user roles against document sensitivity before generating answers.
- Automatically redacts specific sensitive data points while providing the rest of the authorized context.
- Implements a strict "Hard Refusal" protocol to prevent information leaking through inference.
Why this beats prompting it yourself
Manual prompts often fail when users try to "jailbreak" or nudge the AI into revealing clues about restricted data. This skill uses a rigid validation framework that forces a permission check and a structured redaction protocol before any content is output, ensuring compliance is never an afterthought.
Use cases
- Internal HR bots answering policy questions while protecting private employee records.
- Legal department knowledge bases that restrict access to specific case files by counsel level.
- Corporate wikis that need to serve different data to interns versus executives.
Known limitations
Requires the user role and document sensitivity levels to be explicitly provided in the context. It cannot guess permissions for unlabelled data.
Requires explicit role and sensitivity labels in the input context. Cannot verify the truth of provided user roles.
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 2 months ago
Permissions
No special permissions declared or detected