plan-to-tasks
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted plans, specifications, or PRDs and creates multiple Markdown files in the workspace based on that content. This creates a surface for indirect prompt injection.
- Ingestion points: The plan, spec, or PRD is passed to the skill as the primary input argument in SKILL.md.
- Boundary markers: There are no explicit delimiters or instructions to ignore instructions embedded within the provided plan documents.
- Capability inventory: The skill is capable of creating and writing multiple new .md files to the local filesystem as defined in references/original-prompt.md.
- Sanitization: No sanitization or content validation is performed on the ingested plan data.
- [COMMAND_EXECUTION]: The references/canon-maintenance-contract.md file instructs the agent to use local validation scripts located in vault-maintenance/scripts/ for vault and index management.
Audit Metadata