backend-service-implement

Pass

Audited by Gen Agent Trust Hub on Feb 26, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted external files such as API specs and code scaffolds which could contain indirect prompt injections. This is mitigated by a mandatory human-in-the-loop approval step before code generation. Ingestion points: Artifact paths (scaffold, API contract, data model, task breakdown) provided in arguments. Boundary markers: No explicit markers are used in the prompt construction to isolate untrusted content. Capability inventory: The skill has permissions for file writing, configuration editing, and local command execution. Sanitization: No evidence of sanitization for injection patterns within the processed artifacts.
  • [COMMAND_EXECUTION]: The skill executes local vendor scripts (id-generator.sh and index-updater.sh) within the plugin's local directory to manage output folder IDs and documentation indexes. These are legitimate internal administrative operations for the plugin's state management.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 26, 2026, 08:30 PM