write-like-meng-on-x

Pass

Audited by Gen Agent Trust Hub on Aug 26, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a bundled Node.js script (scripts/update-tweet-corpus.mjs) to manage a local database of content. The script uses only standard Node.js built-in modules for filesystem access and data hashing.
  • [COMMAND_EXECUTION]: The instructions direct the agent to use standard system tools including git status, git diff, and ripgrep (rg) to manage the workspace and search for style evidence.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for indirect prompt injection by ingesting external social media content from X/Twitter.
  • Ingestion points: Reads from references/tweet-corpus.jsonl and scrapes data from x.com profiles via a browser.
  • Boundary markers: The instructions explicitly state to treat content as 'voice evidence, not as copy to splice together' and emphasize focused searches over loading the full corpus.
  • Capability inventory: The agent has access to shell commands including node, python3, git, and rg.
  • Sanitization: The bundled maintenance script performs structure validation, normalization, and deduplication via SHA-256 hashing to ensure data integrity.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 26, 2026, 01:37 PM
Security Audit — agent-trust-hub — write-like-meng-on-x