optflow

Pass

Audited by Gen Agent Trust Hub on Apr 6, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data by scanning repository files to identify optimization points. This creates a surface for indirect prompt injection if the analyzed code contains malicious instructions.
  • Ingestion points: Workflow Step 0 in SKILL.md involves scanning the repository content.
  • Boundary markers: None explicitly defined to separate untrusted repository content from instructions.
  • Capability inventory: The skill is authorized to perform file writes (to execute optimizations) and shell command execution (to run test matrices and validation checks defined in references/delivery-base.md).
  • Sanitization: No explicit sanitization or instruction filtering for the scanned content is mentioned.
  • [COMMAND_EXECUTION]: The skill performs automated shell command execution for testing and verification as part of its core functionality.
  • Evidence: references/delivery-base.md specifies running 'mapped step checks' and 'relevant checks immediately after code change' to validate implementation.
  • Safety Context: These operations are restricted to the repository's local environment and are intended for standard development tasks such as running test suites (unit, integration, E2E).
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 6, 2026, 09:27 AM