starworkInit
Pass
Audited by Gen Agent Trust Hub on Nov 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill orchestrates the 'starwork' CLI, specifically using commands like 'starwork init' and 'starwork adapt' to modify the filesystem. These actions are protected by a mandatory 'dry-run' phase and explicit user verification requirements for all target paths and proposed changes.
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes and integrates existing project documentation, such as 'AGENTS.md' and 'CLAUDE.md', which could contain untrusted instructions.
- Ingestion points: Reads existing AI configuration and project documentation files from the user-specified target directory.
- Boundary markers: Utilizes defined content sections like 'Read First' and 'File Boundaries' to structure and delimit instructions.
- Capability inventory: Includes execution of CLI tools with filesystem write permissions and the generation of dynamic initialization blueprints.
- Sanitization: Employs a conservative 'draft-first' approach ('--agent-docs draft') when dealing with non-empty projects to prevent direct overwriting and facilitate review.
Audit Metadata