ts-workflow-implement-review

Pass

Audited by Gen Agent Trust Hub on Aug 4, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting untrusted user data (source requests and developer feedback) and interpolating it into prompts for sub-agents (Implementation and Review Judges).
  • Ingestion points: User-provided content is ingested in the CREATE_LOG (Source request) and HANDLE_DEVELOPER_FEEDBACK (Change requests) sections.
  • Boundary markers: The skill lacks explicit boundary markers or instructions for the sub-agents to ignore embedded commands within the source request or developer feedback fields when they are passed into the judges' task inputs.
  • Capability inventory: The dispatched sub-agents utilize ts-implement and ts-review, which possess capabilities to modify the file system and perform code changes.
  • Sanitization: No sanitization, escaping, or filtering of user-provided strings is performed before they are processed by the sub-agent judges.
  • [COMMAND_EXECUTION]: The skill uses local shell commands to establish a workflow baseline.
  • Evidence: The skill executes git status and git HEAD during the CREATE_LOG phase to record the starting state of the repository.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 4, 2026, 08:47 AM
Security Audit — agent-trust-hub — ts-workflow-implement-review