lit-synthesis

Pass

Audited by Gen Agent Trust Hub on Mar 1, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local shell scripts, specifically scripts/pdf-to-md.sh and scripts/read-paper.sh, to convert PDF documents into Markdown for analysis. These scripts reference a hardcoded path to the docling executable located in the author's local directory (/Users/nealcaren/.local/bin/docling).
  • [EXTERNAL_DOWNLOADS]: The skill documentation recommends the installation of the docling Python package via pip to support PDF conversion functionality.
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface (Category 8) due to its processing of untrusted research papers. \n
  • Ingestion points: Full-text content from PDFs is ingested via the docling conversion process in Phase 1: Deep Reading.\n
  • Boundary markers: The scripts/reading-agent-prompt.md template used for the reading subagent does not include explicit delimiters or instructions to ignore potential commands embedded within the paper text.\n
  • Capability inventory: The skill has the ability to spawn subagents using the Task tool and can perform file-write operations to the local filesystem for generating synthesis reports.\n
  • Sanitization: No sanitization or content filtering is applied to the extracted PDF text before it is analyzed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 1, 2026, 06:46 AM