sketch
Pass
Audited by Gen Agent Trust Hub on Aug 8, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes data from a previous 'interview step' to generate implementation plans. Malicious content in that prior context could attempt to influence the agent's behavior during the file-writing process.
- Ingestion points: The skill explicitly instructs the agent to use the 'chosen approach as settled in the interview step' (SKILL.md).
- Boundary markers: There are no defined delimiters or instructions to ignore embedded commands within the ingested interview data.
- Capability inventory: The skill possesses file-writing capabilities, specifically creating files at
~/.cursor/plans/<name>.plan.md. - Sanitization: The instructions do not define any sanitization, escaping, or validation logic for the input data.
Audit Metadata