social-listening-brief

Pass

Audited by Gen Agent Trust Hub on Jul 3, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The README provides installation instructions for the UnifAPI plugin and MCP server via the /plugin command and npx. These resources are hosted by the skill's author (unifapi-agent).
  • [COMMAND_EXECUTION]: The skill's workflow relies on executing several MCP tools provided by the unifapi service (such as x/tweets/search/recent and tiktok/search) to retrieve live data from social networks.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because its core functionality involves processing untrusted data from public social media posts and comments.
  • Ingestion points: Public content and metadata fetched from X, Reddit, YouTube, TikTok, Threads, Hacker News, and various news sources as described in SKILL.md.
  • Boundary markers: The instructions do not define specific structural delimiters (e.g., XML tags) to separate untrusted platform data from the system prompt, though they do require verbatim quoting for human review.
  • Capability inventory: The skill is designed as a read-only analyst; it lacks capabilities to modify files, execute arbitrary code, or perform network writes based on the ingested content.
  • Sanitization: There are no instructions provided for sanitizing or filtering the content retrieved from external social media APIs.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 3, 2026, 04:06 PM
Security Audit — agent-trust-hub — social-listening-brief