product-glossary

Pass

Audited by Gen Agent Trust Hub on Aug 27, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests content from user-provided Markdown and HTML prototype files which could potentially contain malicious instructions intended to manipulate the analysis performed by subagents.\n
  • Ingestion points: The skill reads docs/product/business-flow.md and all HTML files within the docs/product/demo/ directory.\n
  • Boundary markers: The instructions for subagents lack explicit delimiters or instructions to ignore embedded prompts within the source content.\n
  • Capability inventory: The system performs directory creation (mkdir), file read/write operations for glossary files, and uses the Agent tool to spawn subagents.\n
  • Sanitization: There is no evidence of sanitization or filtering of the source file content before it is passed to the LLM-based subagents for processing.\n- [COMMAND_EXECUTION]: The skill instructions include the use of the shell command mkdir -p docs/product to manage the project's directory structure.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 27, 2026, 02:49 PM
Security Audit — agent-trust-hub — product-glossary