initiative
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided product ideas and project documentation to generate structured initiative artifacts. This ingestion of untrusted external data creates an attack surface for indirect prompt injection, although the skill focuses on data organization rather than execution.
- Ingestion points: User-provided descriptions for initiatives and project documentation files like
project-config.json. - Boundary markers: The skill does not explicitly define sanitization or boundary markers (e.g., XML tags or delimiters) when interpolating input data into templates.
- Capability inventory: The skill can write files to the
initiatives/directory, manage subagents viaspawn_agent, and perform web searches. - Sanitization: No explicit sanitization or filtering of external content is specified in the workflow.
- [COMMAND_EXECUTION]: The skill utilizes environment-scanning tools like
Glob()and task-management tools likespawn_agent. These tools are used for legitimate project management purposes, such as identifying project modules and automating sub-tasks within the local development environment.
Audit Metadata