connections-optimizer
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted third-party content from social media platforms and external research, which introduces an indirect prompt injection surface if profiles or recent posts contain adversarial instructions.
- Ingestion points: External profile text, user handles, recent activity history from X and LinkedIn, and person/company enrichment data fetched via Exa/deep research (specified in
SKILL.md). - Boundary markers: Absent. The instructions do not specify any delimiters or safety boundaries to isolate untrusted social media content from the agent prompt context.
- Capability inventory: Incorporates integration with
x-api, browser automation controls for LinkedIn/X, and local desktop automation for Apple Mail drafting (specified inSKILL.md). - Sanitization: Absent. No filtering, validation, or prompt-escaping mechanisms are described for managing external text input.
Audit Metadata