The Agent Skills Directory

[DATA_EXFILTRATION]: The skill performs network requests to api.acedata.cloud. This domain belongs to the skill's author (acedatacloud) and is the intended endpoint for the unified LLM gateway service.
[PROMPT_INJECTION]: The skill exposes an attack surface for indirect prompt injection as it processes external content through the messages and question parameters.
Ingestion points: User-supplied text in the messages array and the question field in SKILL.md.
Boundary markers: None identified in the provided examples.
Capability inventory: Limited to making network API calls to the vendor endpoint; no access to the local file system or shell execution was detected.
Sanitization: No explicit sanitization or filtering of input is described in the documentation.
[SAFE]: The skill follows security best practices by recommending that API tokens be stored in a .env file rather than being hardcoded.

ai-chat