hackernews-intel
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from the Hacker News Algolia API which is then presented to the agent.
- Ingestion points: Data retrieved from
https://hn.algolia.com/api/v1/search_by_dateinscripts/monitor-hn.js. - Boundary markers: The script outputs story metadata (titles, URLs) to the console with basic prefixes (e.g.,
[NEW]), but lacks strict delimiters to prevent the agent from interpreting title content as instructions. - Capability inventory: The skill can perform network operations (Slack webhook alerts via
fetch) and local file system operations (SQLite database management viabetter-sqlite3). - Sanitization: External content retrieved from Hacker News is not sanitized or escaped before being displayed to the agent or sent to the Slack webhook.
- [SAFE]: The skill uses a vendor-provided installation method via
npxtargeting the author's own domain (opendirectory.dev). This represents normal deployment functionality for this vendor.
Audit Metadata