wizard
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted repository data to determine its code generation logic.
- Ingestion points: Reads .env, README, docker-compose, and GitHub workflow files in the repository to scope wizard stages (SKILL.md Step 1).
- Boundary markers: No delimiters or instructions to ignore embedded instructions within these files are present.
- Capability inventory: Generates bash scripts that perform file writes to .env and set secrets via the gh CLI.
- Sanitization: No sanitization of ingested content is performed before incorporating it into the authored wizard stages.
- [COMMAND_EXECUTION]: The skill authors and prepares bash scripts for execution by the user, providing a surface for potentially malicious commands to be introduced through AI-generated code if the scoping process is compromised.
Audit Metadata