explore
Pass
Audited by Gen Agent Trust Hub on Mar 15, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: Executes a hardcoded bash command using standard utilities like
findandwcto count source files. This operation is limited to assessing the project scale to decide whether to spawn parallel agents. - [PROMPT_INJECTION]: Exhibits a surface for indirect prompt injection. The skill extracts metadata and content from project files (README.md, manifests, source code) and interpolates them into system prompts for sub-agents managed via the
Tasktool. - Ingestion points: Project configuration and source files are read during the orientation phase in
SKILL.md. - Boundary markers: The skill uses Markdown headers to structure agent prompts but does not employ secure delimiters or instructions for sub-agents to ignore potentially malicious content within the ingested data.
- Capability inventory: Sub-agents are granted access to
Bash,Read,Glob,Grep, andTasktools, which can be misused if the agent obeys instructions embedded in the analyzed codebase. - Sanitization: No validation or escaping of the project-derived data is performed prior to its use in prompt templates.
Audit Metadata