vault-skill-factory

Pass

Audited by Gen Agent Trust Hub on Aug 18, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes command-line tools such as qmd, vsearch, grep, and glob to query and retrieve content from the Obsidian vault.
  • [EXTERNAL_DOWNLOADS]: The skill leverages external Python scripts located within the user's local wiki repository (e.g., improve_description.py, package_skill.py) to validate and package the generated artifacts. These are local dependencies intended to be part of the user's workspace.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided wiki pages which could contain untrusted data.
  • Ingestion points: Reads markdown pages from the Obsidian vault via the file system.
  • Boundary markers: The instructions mandate writing output to a dedicated review directory ($SKILL_FACTORY_OUTPUT_DIR) rather than active skill directories, and require the preservation of uncertainty markers like ^[inferred].
  • Capability inventory: Performs file reads, directory creation, and executes local Python scripts.
  • Sanitization: The distillation process transforms raw notes into structured reasoning guides, and the final output requires manual human review and manual installation steps, preventing automated exploitation chains.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 18, 2026, 09:26 PM
Security Audit — agent-trust-hub — vault-skill-factory