audit-before-i-tweet
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-provided tweet drafts which could potentially be used to influence the execution of the
publish_posttool. - Ingestion points: The skill is designed to process tweet drafts provided by users in the chat context, as mentioned in the description and instructions.
- Boundary markers: The instructions lack explicit delimitation or warnings to the AI to ignore instructions embedded within the user-provided draft text.
- Capability inventory: The skill utilizes the
publish_postandimprove_posttools, which perform external write operations to social media platforms. - Sanitization: No sanitization, validation, or escaping of the user-provided draft content is specified before the AI processes it.
Audit Metadata