research

Pass

Audited by Gen Agent Trust Hub on Oct 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions direct the agent to use a command-line utility called story for project maintenance tasks like adding research notes, reindexing, and validation. In cases where the main tool is not globally installed, it utilizes a local fallback script located at ../story-maintenance/scripts/story.js. These operations are limited to project management and do not exhibit malicious behavior.
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to gather and record information from external sources (web searches, user documents, interviews) into the project's markdown files. This creates an attack surface where malicious content in a research source could attempt to influence the agent's context.
  • Ingestion points: External data enters the agent's context through web search results, fetch tools, and documents provided by the user for fact-checking purposes.
  • Boundary markers: Data is organized using standard Markdown headers (e.g., ## Findings, ## Search Plan) and YAML frontmatter fields (e.g., sources, accuracy), providing structural separation.
  • Capability inventory: The agent can perform file writes to the research/ directory and execute project-specific CLI commands for reindexing and word counts.
  • Sanitization: While the skill encourages citations and exact quoting, it does not prescribe specific technical sanitization methods for the text ingested from external sources before it is written to the repository.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 15, 2026, 03:17 PM