community-management

Pass

Audited by Gen Agent Trust Hub on Feb 28, 2026

Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
  • [NO_CODE]: The skill references several local source files (src/leaveAllCommunities.js, src/joinCommunities.js, src/engagementBooster.js, and src/audienceDemographics.js) which are not included in the provided analysis package.
  • [PROMPT_INJECTION]: The documentation describes a process for scraping and acting upon untrusted external data, which introduces a surface for indirect prompt injection.
  • Ingestion points: The joinCommunities.js script scrapes community names and descriptions directly from the X/Twitter discovery page (x.com/i/communities/suggested).
  • Boundary markers: No boundary markers or delimiters are mentioned in the documentation to help distinguish between descriptive metadata and potentially malicious instructions embedded in community names.
  • Capability inventory: The skill is designed to perform automated DOM actions on a social media platform, including clicking Join and Leave buttons and interacting with community members.
  • Sanitization: There is no description of content sanitization or validation of the community strings before they are processed by the keyword matching logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 28, 2026, 09:53 AM