critical-thinking-task-designer

Pass

Audited by Gen Agent Trust Hub on May 8, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it interpolates unvalidated user strings into the core instruction prompt without robust boundary markers.
  • Ingestion points: User-supplied fields in SKILL.md such as curriculum_topic, learner_stage, discipline_or_subject, knowledge_baseline, existing_learning_targets, and thinking_focus.
  • Boundary markers: The prompt uses bold headers to label inputs but lacks formal delimiters (e.g., XML tags or clear isolation markers) to prevent user-supplied text from being interpreted as instructions by the underlying model.
  • Capability inventory: The skill does not possess any dangerous capabilities; it has no file system access, network connectivity, or subprocess execution tools across any of its components.
  • Sanitization: There is no evidence of input validation, escaping, or sanitization performed on the user-provided data before it is inserted into the prompt template.
Audit Metadata
Risk Level
SAFE
Analyzed
May 8, 2026, 03:19 AM