create-implementation-plan
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest untrusted data from a user to generate a machine-readable implementation plan. This plan includes autonomous verification commands and task descriptions that are intended to be executed by subsequent AI agents or humans, creating a vulnerability surface where malicious instructions could be embedded in the requirements.
- Ingestion points: The skill takes input through the
${input:PlanPurpose}variable and through conversation via theAskUserQuestiontool. - Boundary markers: Absent. The instructions do not define delimiters or provide warnings to ignore embedded instructions within the user-provided scope or requirements.
- Capability inventory: The skill itself performs file-write operations to the
/plan/directory. The generated output contains aVerificationblock designed for autonomous command execution by downstream processes. - Sanitization: Absent. There are no instructions to sanitize, escape, or validate the user-provided requirements before they are incorporated into the executable plan structure.
Audit Metadata