workflow-quality-gate

Pass

Audited by Gen Agent Trust Hub on Oct 7, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow where the agent ingests diagnostic data from multiple security and testing skills (e.g., test-red-team, audit-security). This configuration is susceptible to indirect prompt injection, as malicious instructions embedded in the audited source code could be processed by the tools and subsequently influence the agent's decision-making or its 'inline fix' capabilities.
  • Ingestion points: SKILL.md instructions consume outputs from various diagnostic gates including test-red-team, audit-security, and test-unit.
  • Boundary markers: Absent; the instructions do not include specific delimiters or directives to ignore instructions found within the diagnostic tool outputs.
  • Capability inventory: The skill references the ability of the audit-security component to perform 'inline' fixes, which implies file-write permissions.
  • Sanitization: Absent; no input sanitization or validation protocols are specified for the tool results before they are processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 7, 2026, 11:27 PM