lit-review

Pass

Audited by Gen Agent Trust Hub on Sep 18, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data when reading paper drafts and bibliographies during initial setup and when ingesting content from web searches to find new literature.
  • Ingestion points: SKILL.md sections Mode C4 (Reading source documents) and Mode B3 (Search for new papers).
  • Boundary markers: The instructions do not specify the use of clear delimiters or warnings to ignore embedded instructions when processing external research content.
  • Capability inventory: The skill has capabilities to perform shell commands (mkdir), write files to the user's filesystem, and perform web searches.
  • Sanitization: While the skill attempts to avoid certain binary formats (.dta, .csv, .sav), it does not describe methods for sanitizing the natural language content of processing academic documents.
  • [COMMAND_EXECUTION]: The skill uses shell commands for project directory management.
  • Evidence: Execution of mkdir -p in Mode S and Mode C to create configuration and review project folders. These are restricted to organizing the workspace.
  • [EXTERNAL_DOWNLOADS]: The skill constructs external links to assist the user in locating and downloading research materials.
  • Evidence: Generation of Google Scholar search URLs and DOI resolution links to help retrieve papers missing from the local repository. These target well-known academic domains (scholar.google.com and doi.org).
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 18, 2026, 08:52 AM
Security Audit — agent-trust-hub — lit-review