website-launch-gate
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data which could contain malicious instructions designed to subvert the agent's behavior.
- Ingestion points: The skill instructions in
SKILL.md(Workflow Step 2) direct the agent to "inspect both the source and the rendered website where possible." - Boundary markers: The skill includes some structural boundaries, such as requiring explicit user authorization before altering production data or deploying (Step 1) and mandating a specific tabular output format with evidence-based status markers (Output section).
- Capability inventory: The agent is empowered to identify and potentially run repository build and test commands (Step 1) and is instructed to "Fix in-scope failures" by modifying files if requested by the user (Step 4).
- Sanitization: The instructions do not define any specific sanitization, filtering, or escaping mechanisms for the external website content being processed.
Audit Metadata