tutor-setup
Pass
Audited by Gen Agent Trust Hub on Apr 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to install system utilities using standard package managers (
brew install popplerorapt-get install poppler-utils). - [COMMAND_EXECUTION]: Uses the
Bashtool to executepdftotextfor PDF processing. The command writes output to/tmp/source.txt, which is a system-wide directory outside the skill's stated 'CWD Boundary Rule'. - [EXTERNAL_DOWNLOADS]: Fetches and installs software packages from well-known system registries during setup.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection from untrusted source documents and code.
- Ingestion points: Reads files via
ReadandGlob; fetches web content viaWebFetch; converts PDFs viapdftotext(Bash). - Boundary markers: Includes a rule to restrict file operations to the current working directory, but lacks delimiters or instructions to ignore embedded agent commands within the ingested content.
- Capability inventory: The agent has access to
Bash,Write,Edit, andWebFetchtools, providing a path for potential exploitation of injected instructions. - Sanitization: There is no description of input validation or content filtering to prevent the processing of malicious prompts embedded in source materials.
Audit Metadata