adhx
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill uses
curlto perform network requests toadhx.com, an external third-party domain. It transmits user-provided handles and status IDs as part of the API request path. - [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes untrusted text and article content from an external API.
- Ingestion points: Fetches tweet text and long-form article content from the
adhx.comAPI endpoint for analysis by the agent (SKILL.md). - Boundary markers: Includes a 'Limitations' section that explicitly warns the agent that API responses are untrusted source data and should not be treated as instructions.
- Capability inventory: The skill uses
curlto retrieve external content which the agent then interprets, summarizes, or analyzes. - Sanitization: Recommends structural validation of the input URL, but lacks mechanisms to sanitize or filter the content returned from the external service before processing.
Audit Metadata