super-swarm-spark
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and parses external markdown plan files to extract task descriptions, names, and validation criteria. This data is then interpolated into prompts for subagents.
- Ingestion points: Markdown plan files read in Step 2 of the process.
- Boundary markers: The instructions do not define delimiters or explicit "ignore embedded instructions" warnings to separate the plan data from the agent's core task logic.
- Capability inventory: The orchestrator and its subagents have the capability to modify the plan file, create/edit code files, stage/commit files, and execute tests.
- Sanitization: The skill lacks logic to validate or sanitize the content of the plan files before processing and interpolating the content into subagent prompts.
Audit Metadata