AI Agent Prompts
Agent Permission Boundary Designer
Translate an agent's responsibilities into least-privilege tool permissions, scoped credentials, approval rules, and revocation controls.
FreeClaudeChatGPT
Best for
Security and platform teams granting agents enough access to work while minimizing the damage from mistakes or manipulation.
Suitable LLM groups
FrontierReasoning
Prompt
You are an AI security architect who specializes in least-privilege access for autonomous agents.
Inputs:
1. Agent responsibilities and prohibited outcomes: {{responsibilities}}
2. Systems, tools, and data involved: {{systems_and_data}}
3. Required read, write, execute, and communication actions: {{required_actions}}
4. Identity, authentication, and approval options: {{security_options}}
5. Compliance, audit, and operational constraints: {{constraints}}
Do the following:
1. Map each responsibility to the minimum data objects, operations, environments, and time windows required; flag permissions that are convenient but not necessary.
2. Design permission tiers for routine, elevated, and emergency actions, including scoped credentials, just-in-time access, approval, rate limits, and transaction limits.
3. Define controls against confused-deputy behavior, indirect prompt injection, cross-user data leakage, credential exposure, and unauthorized tool chaining.
4. Specify denial behavior, escalation, credential rotation, revocation, break-glass access, and immutable audit events.
5. Output a permission matrix, trust-boundary diagram in text, sample policy rules, abuse-case tests, and a rollout sequence that starts with the safest viable scope.How to use
- List actions at the API-operation level where possible.
- Separate required permissions from convenient access.
- State monetary, temporal, and volume limits.
- Have security owners review the resulting permission matrix.
Example input
Responsibilities: Resolve tier-1 ecommerce support requests, issue refunds under EUR 75, and update shipping addresses before dispatch; never view full card data or cancel fulfilled orders. Systems: help desk, order database, shipping portal, and payment processor. Actions: read customer and order records, draft replies, change unshipped addresses, and issue limited refunds. Security options: service identities, OAuth scopes, approval queue, and per-action logs. Constraints: GDPR, 24/7 operation, 30 refunds per hour maximum, and immediate revocation by an on-call engineer.
Example output
The matrix grants order.read only for the active ticket's customer, shipping.address.write only before the dispatch flag, and refund.create with a EUR 75 per-transaction and 30-per-hour cap. The payment processor returns masked instruments only. Refunds above EUR 75 require a time-limited elevated token approved by a supervisor. Tool outputs are treated as untrusted content, and no retrieved text can request broader scopes. The design includes denial messages, revocation events, and abuse tests for cross-customer access and chained refund attempts.
Customization tips
- — Add tenant, record, and field-level data scopes.
- — Use short-lived credentials instead of broad permanent tokens.
- — Define a safe read-only degradation mode.
- — Include real abuse cases from threat modeling.
Tags
#least-privilege#agent-security#access-control#tool-permissions#identity-management
FAQ
What is this prompt for?
It converts an agent's job into an enforceable least-privilege access and credential design.
How should I customize it?
Provide specific systems, objects, API operations, transaction limits, identities, and prohibited outcomes rather than broad tool names.
Are there any limitations?
The generated policy is an architecture proposal, not a substitute for platform-specific security testing, legal review, or formal threat modeling.
How is it different from a basic prompt?
It maps every responsibility to minimum permissions and adds tiering, just-in-time elevation, denial behavior, revocation, audit events, and abuse tests.
Related Prompts
All AI Agents prompts →Free
Agent Deployment Readiness Reviewer
Run a structured pre-production review across product value, evaluation, security, operations, fallbacks, and rollout controls.
ClaudeChatGPT
#agent-deployment#production-readiness#launch-review
Free
Human-in-the-Loop Checkpoint Designer
Place human review at the right moments in an agent workflow based on risk, reversibility, confidence, and operational cost.
ClaudeChatGPT
#human-in-the-loop#agent-governance#approval-workflow
Free
Tool Definition Writer for Agents
Writes a clear, well-scoped tool definition (name, description, parameters, and usage guidance) for an AI agent, designed to minimize misuse and ambiguous tool-calling.
ClaudeChatGPT
#ai-agents#tool-definitions#function-calling
Related Skills
Safety, Privacy & ComplianceFree
Data Access and Least Privilege Reviewer
Review data access for least privilege, role design, sensitive fields, segregation, approvals, logging, reviews, and revocation.
ClaudeChatGPT
#least privilege#data access review#privacy controls
Agent Systems & LLM WorkflowsFree
Human in the Loop Agent Control Architect
Design human-control patterns for AI agents with approval gates, previews, authority limits, interruption, rollback, escalation, and auditability.
ClaudeChatGPT
#human in the loop#agent controls#AI governance
Related articles coming soon — visit the Learn hub.