spark-optimization
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources and local files to generate optimization artifacts, creating a potential surface for indirect instructions.
- Ingestion points: External datasets or model inference results retrieved via
sandbase_run(SKILL.md, references/sandbase-api-map.md), and local input files (SKILL.md Workflow step 1). - Boundary markers: The skill requires restating acceptance checks and constraints before execution (SKILL.md Workflow step 2).
- Capability inventory: Generation of Spark implementations, artifact production, and runtime diagnostics.
- Sanitization: No explicit sanitization or filtering of external content is described.
Audit Metadata