skills.rest
FFOO6866
FFOO6866COMMUNITY
Explore all skills by this author →
GitHub

29-pact

Enforce fail-closed PACT governance for AI agent actions with audit verdicts.

Updated Oct 10, 2025
Search Tags:#audit trail#agent authorization#pact governance#envelope constraints#knowledge clearance#dtr addressing
One-click install
npx skills add https://github.com/FFOO6866/lead2cash --skill 29-pact-ffoo6866
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: 29-pact
Source: https://github.com/FFOO6866/lead2cash/tree/main/.claude/skills/29-pact
Command: npx skills add https://github.com/FFOO6866/lead2cash --skill 29-pact-ffoo6866

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps you design and apply PACT governance controls for AI agents, including D/T/R accountability addressing, constraint envelopes, knowledge clearance, and default-deny access enforcement.

Core Features & Use Cases

  • GovernanceEngine decisioning: Compute effective constraint envelopes, evaluate actions, run multi-level verification, and produce audit-ready verdicts.
  • Fail-closed access control: Use a 5-step clearance-and-containment algorithm (DEFAULT IS DENY) with KSPs and Pact Bridges for scoped sharing.
  • Governed agents & middleware: Wrap agents to enforce policy before tool execution and integrate governance checks into custom frameworks (including MCP).
  • Monotonic tightening & TOCTOU defenses: Apply envelope intersection with strict tightening validation and include version hashes in verdicts to prevent stale-snapshot misuse.
  • Conformance and observability: Support normative conformance vectors (N1–N6), tiered audit durability, and structured observation sinks for monitoring.

Quick Start

Run the pact-governance-engine skill to set up a GovernanceEngine, attach envelopes, and call verify_action for a first governed decision.

Frequently Asked Questions about 29-pact

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I enforce fail-closed access control for AI agent tool execution?▼

Fail-closed access control for AI agents is enforced by applying a 5-step clearance-and-containment algorithm that defaults to deny, using KSPs and Pact Bridges for scoped sharing before any governed tool execution occurs.

What is PACT governance and how does D/T/R addressing work for agent accountability?▼

PACT governance enforces deterministic constraint envelopes and fail-closed action control for AI agents across D/T/R addressed organizations, ensuring strict accountability and default-deny access before tool execution.

How do I prevent TOCTOU attacks during AI agent knowledge clearance checks?▼

Prevent TOCTOU attacks during knowledge clearance checks by applying envelope intersection with monotonic tightening validation and including TOCTOU envelope version hashes in governance verdicts to block stale-snapshot misuse.

Can I integrate PACT governance checks into custom agent frameworks and MCP tool calls?▼

You can integrate PACT governance into custom agent frameworks and MCP tool calls by wrapping agents with middleware that enforces policy and evaluates actions before executing governed MCP tool calls.

How do I set up a GovernanceEngine to compute effective constraint envelopes?▼

Set up a GovernanceEngine by running the pact-governance-engine skill, attaching YAML-described constraint envelopes, and calling verify_action to compute multi-level verification and produce audit-ready verdicts.

What is the best way to generate audit-ready governance verdicts for AI agents?▼

Generate audit-ready governance verdicts by using a GovernanceEngine to evaluate actions, run multi-level verification, and output structured observation sinks supporting normative conformance vectors N1–N6 and tiered audit durability.

Related Skills

Explore More Software Engineering Skills→
esperieesperie

29-pact

Implement configurable envelopes, approval workflows, and tamper-evident audit chains for AI agent governance.

Community
Advanced
gonzalezpazmonicagonzalezpazmonica
47

governance-enterprise

Orchestrate enterprise governance workflows for audit trails, compliance, and certifications.

Community
Advanced
msaad00msaad00
29

agent-bom-enforce

Intercept and enforce user-defined policies on MCP tool calls via a runtime proxy.

Community
Advanced
LaunchPadLabLaunchPadLab
1

policy-patterns

Implements POMDP-based reward shaping for LLM agents with formal guarantees.

Official
Intermediate