wasm-spa-autofix-react-imports

Pass

Audited by Gen Agent Trust Hub on Feb 25, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits a potential surface for indirect prompt injection.
  • Ingestion points: Untrusted data enters the agent context via the fileContents and bundlerLogs inputs.
  • Boundary markers: The skill logic lacks explicit instructions or markers to differentiate between the agent's instructions and potentially malicious content embedded in the analyzed code or logs.
  • Capability inventory: The skill is capable of generating code patches (returned in the patches output field) which are intended to be applied directly to the project's source code.
  • Sanitization: No specific sanitization or validation of the input data is mentioned before it is used to inform the generation of code patches.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 25, 2026, 01:05 AM