write-like-meng-on-x
Pass
Audited by Gen Agent Trust Hub on Jul 28, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes a bundled Node.js script (update-tweet-corpus.mjs) for corpus ingestion and deduplication. It also runs a validation utility using python3.
- [EXTERNAL_DOWNLOADS]: The skill is instructed to retrieve public authored posts from X (Twitter) to update the style references stored in references/tweet-corpus.jsonl.
- [PROMPT_INJECTION]: The skill processes external social media content, creating a surface for indirect prompt injection.
- Ingestion points: Data is sourced from references/tweet-corpus.jsonl and external x.com status pages.
- Boundary markers: Instructions explicitly direct the agent to treat these posts as voice evidence, preventing the interpretation of ingested content as direct instructions.
- Capability inventory: The skill uses Bash, Write, and Edit tools to maintain the local database.
- Sanitization: The ingestion script normalizes text and uses cryptographic hashing to identify and prevent duplicate content entry.
Audit Metadata