dspy-optimization
Pass
Audited by Gen Agent Trust Hub on Feb 24, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes external training data which presents a surface for indirect prompt injection. Malicious instructions embedded in training examples could influence the behavior of the optimized programs.
- Ingestion points: The
scripts/optimize-dspy.pyscript loads training examples from a user-provided JSONL file. - Boundary markers: No delimiters or 'ignore instructions' warnings are present in the data loading process.
- Capability inventory: The script can write model artifacts to disk using
optimized.save(). - Sanitization: No input validation or sanitization is performed on the training data content.
Audit Metadata