skill-from-book

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • PROMPT_INJECTION (LOW): The skill is designed to ingest and process text from external sources (e.g., book.md), which represents a surface for indirect prompt injection. Malicious instructions within the source book could potentially influence the behavior of the extraction subagents.\n
  • Ingestion points: references/extraction-patterns.md defines subagent tasks that read source book content.\n
  • Boundary markers: While task instructions define specific line ranges, the prompt templates lack explicit delimiters or instructions to ignore embedded commands within the source text.\n
  • Capability inventory: The skill includes filesystem write capabilities via scripts/init-book-skill.py and suggests shell command execution in references/analysis-guide.md.\n
  • Sanitization: No input validation or sanitization of the source material is mentioned in the guides or the utility script.\n- COMMAND_EXECUTION (SAFE): The scripts/init-book-skill.py utility and the shell commands recommended in references/analysis-guide.md (such as wc and grep) are benign. The Python script uses standard libraries (argparse, os, pathlib) to create a project structure and write boilerplate text without any unsafe execution patterns or external dependencies.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:07 PM