graph-evolution

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted source code from git snapshots to compute structural differences. \n
  • Ingestion points: Source code snapshots accessed via git worktrees. \n
  • Boundary markers: The skill provides a structured markdown report template and a quality checklist that includes human review requirements. \n
  • Capability inventory: Employs git worktree, uv run, and Python execution to perform graph analysis. \n
  • Sanitization: Relies on structural metrics (cyclomatic complexity, blast radius) and graph relationships rather than direct execution of code content.\n- [DYNAMIC_EXECUTION]: The skill provides Python code snippets for the agent to build code graphs and run pre-analysis via the trailmark API. \n- [COMMAND_EXECUTION]: Uses standard git commands and temporary directory utilities to manage snapshots for comparison. \n- [EXTERNAL_DOWNLOADS]: Instructs the installation of the trailmark tool, which is a vendor-owned analysis resource.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 01:17 PM
Security Audit — agent-trust-hub — graph-evolution