comfy-build
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill acknowledges the risk of processing attacker-controlled data from external sources such as build logs and custom node package metadata.\n
- Ingestion points: Processes logs from
comfy build release logsand scans local file system metadata during environment initialization.\n - Boundary markers: Explicitly instructs the agent to treat build logs as "attacker-controlled text" and forbids executing any commands or copying literals found within them.\n
- Capability inventory: Access to
comfy-clifor environment scanning, configuration management, and platform synchronization.\n - Sanitization: Requires the agent to describe findings in its own words and mandates explicit user confirmation before executing potentially destructive or billable actions.
Audit Metadata