lead-generation
Pass
Audited by Gen Agent Trust Hub on Feb 28, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: Vulnerability to indirect prompt injection through social media data.
- Ingestion points: Untrusted data enters via x_search_tweets, x_get_profile, and x_get_tweets from external user accounts.
- Boundary markers: No delimiters or specific instructions are used to separate external content (tweets, bios) from the agent's logic.
- Capability inventory: The skill uses MCP tools for account interaction and references scripts for automated engagement and following.
- Sanitization: No evidence of sanitization for tweet text or bio content before processing for intent scoring.
- [COMMAND_EXECUTION]: Reference to unverified automation scripts.
- Evidence: Mentions files like src/automation/keywordFollow.js and src/engagementBooster.js.
- Risk: The behavior of these scripts cannot be audited as they are external to the analyzed file and perform automated interactions on social platforms.
Audit Metadata