soia-dev-archify-diagrams
Pass
Audited by Gen Agent Trust Hub on Aug 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill executes external binaries via
spawnSyncto perform diagram rendering and screenshot capture.\n - Evidence:
render-archify-diagrams.mjscallsarchify.mjsandexport-archify-previews.mjscallschrome.\n- [EXTERNAL_DOWNLOADS]: TheSKILL.mdfile points to an external third-party repository (github.com/tt-a1i/archify.git) for its core rendering dependency and provides instructions for manual user installation.\n- [PROMPT_INJECTION]: The skill ingests user-supplied JSON files (e.g.,.architecture.json,.workflow.json) which are processed by local scripts and external tools, creating a surface for indirect prompt injection.\n - Ingestion points:
scripts/render-archify-diagrams.mjs\n - Boundary markers: Absent\n
- Capability inventory: Command execution (spawnSync) and file system writes (fs.writeFile)\n
- Sanitization: Absent; content is passed to the rendering engine.\n- [DATA_EXFILTRATION]: The skill reads configuration files from
~/.config/soia-skills/to locate required tools and define output directories, which is a standard procedure for maintaining tool settings and does not involve unauthorized data movement.
Audit Metadata