ai-festival-poster
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The
scripts/get_fonts.pyscript downloads font files from the official Google Fonts GitHub repository (google/fonts). This is a trusted, well-known service for open-source assets. - [COMMAND_EXECUTION]: The skill requires running local Python scripts to scan session logs and render images. The procedural painting logic in
scripts/pp/uses NumPy and SciPy to generate visuals locally without external image model APIs. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted session data from local logs.
- Ingestion points:
scripts/scan_usage.pyreads session history files from~/.claude/projects/and~/.codex/sessions/. - Boundary markers: The
SKILL.mdfile contains explicit privacy rules instructing the agent to report only aggregated counts, never paste raw log contents into the chat, and perform a manual privacy check with the user. - Capability inventory: Filesystem read access for logs, local JSON writing for usage statistics, and image rendering via Pillow.
- Sanitization: The scanning process is limited to counting occurrences of specific tool and model identifiers using regular expressions, minimizing the risk that malicious text within logs could trigger unauthorized actions.
Audit Metadata