content-digest

Pass

Audited by Gen Agent Trust Hub on Apr 7, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill processes arbitrary data from external URLs and local files to generate content digests, which introduces an indirect prompt injection surface. Content ingested via WebFetch or Read could contain instructions intended to manipulate the agent's summarization behavior. Ingestion points: WebFetch and Read tools (SKILL.md). Boundary markers: None identified. Capability inventory: Bash script execution, web fetching, file system access, and note management (SKILL.md). Sanitization: None implemented.\n- [COMMAND_EXECUTION]: The skill utilizes local shell commands to handle file format conversions, specifically using the macOS sips utility for image formatting and a local Python script (pdf_to_images.py) to render PDF pages. These commands are integral to the skill's stated purpose of multi-format content analysis.\n- [EXTERNAL_DOWNLOADS]: The included PDF processing script automatically installs the pymupdf library from the Python Package Index (PyPI) at runtime if missing. While PyPI is a standard and trusted package registry, this dynamic installation represents an external dependency download and execution at runtime.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 7, 2026, 04:00 AM
Security Audit — agent-trust-hub — content-digest