sdd-apply

Pass

Audited by Gen Agent Trust Hub on Mar 14, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The agent is instructed to identify and execute test runners based on project configuration files like package.json, pyproject.toml, or a custom config.yaml. This is a core functional requirement for the implementation cycle to verify code changes.\n- [PROMPT_INJECTION]: The skill loads instructions from external project artifacts and a skill registry, creating a surface for indirect prompt injection if those files contain adversarial content.\n
  • Ingestion points: Reads instructions from tasks.md, .atl/skill-registry.md, and architectural artifacts (spec, design, proposal) retrieved from the memory store.\n
  • Boundary markers: The instructions do not specify delimiters or constraints for the content of these external files.\n
  • Capability inventory: The agent can perform memory updates, save new observations, and execute shell commands identified as test runners.\n
  • Sanitization: There is no documented logic for sanitizing or filtering instructions retrieved from external project-controlled files before they are followed.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 14, 2026, 02:21 AM