What problem does it solve?
AI agents, like human developers, can inadvertently deviate from established task protocols, leading to critical violations (e.g., skipping user approvals, main agent directly implementing code) that compromise quality, security, or workflow integrity. Manually auditing complex conversations for these subtle violations is time-consuming, subjective, and prone to oversight. This Skill automates the compliance check.
Core Features & Use Cases
- Systematic Compliance Checks: Automatically audits conversation sessions against a comprehensive set of protocol rules (e.g., user approval, worktree isolation, agent delegation, state transitions).
- Evidence-Based Reporting: Provides clear, objective verdicts for each check, citing specific timeline evidence (timestamps, tool uses, actors) for any violations.
- Documentation Fix Recommendations: Suggests concrete changes to protocol documentation to prevent future recurrences of identified violations, fostering continuous improvement.
- Use Case: After an agent completes a task, you need to verify it followed all mandatory steps, such as getting user approval after the SYNTHESIS phase. This Skill automatically scans the conversation, flags any missed approvals as CRITICAL violations, and recommends updating the protocol documentation to reinforce the rule, ensuring "AI works, you rest."
Quick Start
Audit the current session for all task protocol compliance violations and recommend documentation fixes.
