The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains 'Agent Interaction Rules' in SKILL.md that use high-priority directives like 'CRITICAL', 'NEVER', and 'MUST' to restrict the agent's output behavior, specifically forbidding the inclusion of code snippets in responses and enforcing 'GET-only' execution for the agent's own calls. These are instructions designed to override the agent's default operational behavior.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes real-time data from an external WebSocket stream, which could potentially be used to deliver malicious instructions to the agent during summarization.
Ingestion points: Incoming WebSocket messages are captured in the onmessage handler of the connectSpotWsAccount function in SKILL.md.
Boundary markers: Absent; the instructions do not define specific delimiters or 'ignore' warnings for the agent to use when processing the external data payloads.
Capability inventory: The skill provides logic for HTTP POST, PUT, and DELETE requests (to manage Listen Keys) and persistent WebSocket connections.
Sanitization: While the skill requires validation for user-provided parameters, it does not specify sanitization or filtering for the content pushed from the server-side WebSocket events.
[EXTERNAL_DOWNLOADS]: The TypeScript code provided in SKILL.md depends on the pako Node.js library to decompress binary WebSocket messages.

bingx-spot-ws-account