skills/hazat/pi-config/plan/Gen Agent Trust Hub

plan

Pass

Audited by Gen Agent Trust Hub on Aug 25, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local shell commands and tool-specific instructions ('sc', 'git', 'ls', 'find', 'rg') to perform repository discovery, state management, and orchestration of sub-agents.
  • [DYNAMIC_EXECUTION]: The skill dynamically generates prompts and configuration files for sub-agents based on the plan and execution state. It employs secure handling of user-derived content by using 'jq' to serialize data into JSON files, mitigating potential command injection risks during agent dispatch.
  • [INDIRECT_PROMPT_INJECTION]: The skill reads and processes codebase content and tool outputs, creating a surface for indirect prompt injection. This is mitigated through a structured evidence chain:
  • Ingestion points: Repository discovery tools (ls, find, rg), plan artifacts, and tool outputs.
  • Boundary markers: The workflow includes a mandatory human-in-the-loop 'Final Checkpoint' where the user must approve the plan and approach before execution begins.
  • Capability inventory: Access to file system reads/writes, shell command execution, and sub-agent orchestration.
  • Sanitization: Strict verification of worker outputs, including Git commit SHA validation and manual reconciliation of artifact contents against the original plan.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 25, 2026, 06:34 AM
Security Audit — agent-trust-hub — plan