suggesting-cursor-rules
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is designed to manage local configuration within the
.cursor/rules/directory. This is a standard and expected behavior for developer tools and does not involve unauthorized network access, sensitive data exposure, or execution of untrusted code. - [INDIRECT_PROMPT_INJECTION]: The skill uses conversation history and user feedback as input to generate project rules, creating a theoretical attack surface for indirect prompt injection.
- Ingestion points: The skill triggers based on repeated patterns in user corrections and natural language feedback as described in the 'Triggers' section of SKILL.md.
- Boundary markers: There are no explicit instructions to the agent to filter or ignore potentially malicious prompt instructions when drafting the rule content.
- Capability inventory: The skill utilizes the agent's capability to create and write markdown files in the local filesystem (
.cursor/rules/*.mdc). - Sanitization: The skill relies on natural language confirmation from the user ('Want me to create a Cursor rule...?') rather than programmatic sanitization or strict schema validation.
Audit Metadata