skills/cometworks/skills/se-dev/Gen Agent Trust Hub

se-dev

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The preparation scripts (GraphifyPrepare.bat, graphify-prepare.sh) automate the installation of the graphifyy package from PyPI, an official and well-known package registry, using standard package managers such as uv, pipx, or pip.
  • [COMMAND_EXECUTION]: The skill executes various system commands for environment setup and maintenance, including disk space verification via PowerShell on Windows and directory management (mkdir, rmdir, rm -rf) across supported platforms.
  • [DYNAMIC_EXECUTION]: The health check utility (graphify-check.sh) executes dynamically generated Python code using heredocs to validate the integrity and community structure of JSON clustering data produced by the graphing tool.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests user-supplied project code (scripts, mods, and plugins) to build navigable structural graphs. It includes mitigation measures such as an explicit exclusion list for non-code files (e.g., .md, .txt, .pdf) to prevent accidental processing of unstructured text by LLM-based extraction backends.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 09:17 AM
Security Audit — agent-trust-hub — se-dev