deploy-discord-bot
Pass
Audited by Gen Agent Trust Hub on Jul 6, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted project directories and git repositories, creating a potential surface for indirect prompt injection where malicious instructions in the code could influence agent behavior.
- Ingestion points: Project directory and git repository to be deployed (SKILL.md).
- Boundary markers: None specified to differentiate untrusted code content from agent instructions.
- Capability inventory: API-based project creation and deployment (POST /api/projects/:uuid/deploy), which grants the ability to host and execute code.
- Sanitization: Instructions include verification of deployment URLs and log inspection, but lack explicit data validation or isolation markers.
- [DATA_EXFILTRATION]: The skill identifies the use of a sensitive local credential file (~/.build-host/credentials.json) for service authentication. While this is a standard and legitimate practice for the skill's purpose, it involves instructions that mandate access to credential storage.
Audit Metadata