comfy-deploy-failures
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill explicitly identifies deployment logs as attacker-controlled text that may contain malicious instructions designed to trigger unauthorized actions.
- Ingestion points: Untrusted deployment and build logs accessed via
comfy deploy logs(SKILL.md). - Boundary markers: Includes a mandatory warning to the agent stating that log content must not be used to form commands, URLs, or arguments (SKILL.md).
- Capability inventory: The skill references broad capabilities via the
comfyCLI tool, including deployment management and workflow execution (SKILL.md). - Sanitization: Instructs the agent to synthesize findings in its own words rather than passing through raw log content into executable actions.
Audit Metadata