project-setup
Pass
Audited by Gen Agent Trust Hub on Jun 12, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill implements an interactive discovery interview (Phase 1) where user-provided input is used to generate and write configuration files such as
CLAUDE.md,.claude/rules/, andGEMINI.md. This creates a surface for indirect prompt injection where malicious or poorly formatted user responses could influence the future behavior of agents operating within the project. - Ingestion points: User input from the Phase 1 Discovery Interview is ingested and written to the file system in Phase 3.
- Boundary markers: The instructions do not specify the use of delimiters (e.g., XML tags or backticks) to wrap user-provided data when generating configuration files.
- Capability inventory: The skill uses the
Writetool to modify the local file system andBashfor operational tasks. - Sanitization: There are no instructions provided to sanitize, validate, or escape the user's input before it is interpolated into the configuration file templates.
Audit Metadata