document-inventory
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through the processing of untrusted document metadata.
- Ingestion points: Extraction of document properties (Title, Author, Subject, etc.) from files found in the workspace as described in SKILL.md.
- Boundary markers: The output format lacks explicit markers or instructions to isolate extracted metadata from the agent's execution context.
- Capability inventory: The skill provides command templates for Bash and PowerShell to perform file operations and Git interactions.
- Sanitization: There is no mention of sanitizing or validating document metadata before it is passed to the orchestrating agent.
- [COMMAND_EXECUTION]: The skill provides shell command templates (find, git, PowerShell) for discovery and inventory tasks. These templates rely on external input for paths and tags, which could lead to command injection if the calling agent does not properly validate those inputs.
Audit Metadata