find-literature

Pass

Audited by Gen Agent Trust Hub on May 18, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads and installs the asta-plugins package from the official Allen Institute for AI GitHub repository. This is a verified vendor resource required for the skill to function.
  • [COMMAND_EXECUTION]: The skill uses the asta CLI for academic searches and jq for processing results. These commands are restricted to the intended use case of literature discovery and metadata extraction.
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by ingesting abstracts and snippets from external academic papers.
  • Ingestion points: Output files from asta literature find and asta papers stored in the .asta/documents/ directory.
  • Boundary markers: The use of structured JSON fields for titles, abstracts, and snippets provides natural boundaries between data and agent instructions.
  • Capability inventory: The skill is limited to the asta CLI toolset, jq data manipulation, and basic file operations like mkdir and mv.
  • Sanitization: Data is handled through structured JSON parsing and jq filtering, which prevents content from being interpreted as command instructions during the retrieval process.
Audit Metadata
Risk Level
SAFE
Analyzed
May 18, 2026, 03:43 AM
Security Audit — agent-trust-hub — find-literature