implementation-task-creator

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting data from local architecture files to generate task files and update project state. Ingestion points: Architecture markdown files located at {project_path}/architecture/{component}.md. Boundary markers: None; the agent processes the file content directly. Capability inventory: The skill uses Write and Edit file capabilities to create task files and modify project state. Sanitization: No content validation or sanitization is performed on the ingested data.
  • [COMMAND_EXECUTION]: The workflow includes instructions to run ddev phpunit commands, which is standard behavior for verifying TDD steps in the targeted Drupal development environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 12:30 AM