aibtc-news-correspondent
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local TypeScript scripts using
bun runfor core operations such as claiming beats and filing signals (e.g.,aibtc-news/aibtc-news.ts). It also utilizes the GitHub CLI (gh api graphql) to interact with and post to GitHub Discussions. - [EXTERNAL_DOWNLOADS]: Fetches live market data and Bitcoin fee information using
curlfrom well-known services, specificallyapi.coinbase.comandmempool.space. - [INDIRECT_PROMPT_INJECTION]: The skill's research workflow involves ingesting data from untrusted external sources, including X.com (via Grok API), ecosystem activity (Moltbook), and GitHub protocol releases. This content could contain instructions designed to manipulate the agent's reporting output.
- Ingestion points: Research process reads from social media, GitHub, and general web sources (SKILL.md, Step 3).
- Boundary markers: There are no explicit instructions or delimiters provided to the agent to distinguish between data and instructions within these external sources.
- Capability inventory: The agent has the ability to file news signals to a remote endpoint and create discussions on GitHub.
- Sanitization: The skill lacks specified sanitization or filtering logic for the data retrieved from external research before it is incorporated into final signals.
Audit Metadata