The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill implements a self-update and modular loading system that fetches content from Bybit's official API and GitHub repository.
Fetches manifest from api.bybit.com.
Downloads module content from raw.githubusercontent.com/bybit-exchange/skills/main/.
Employs SHA256 integrity verification for all downloaded files before execution or installation.
[COMMAND_EXECUTION]: Several shell-based operations are used for core functionality.
Uses curl for network requests to the Bybit API and GitHub.
Uses openssl for HMAC-SHA256 and RSA-SHA256 request signing.
Uses shasum and awk for file integrity verification during updates.
Uses node to execute modular JavaScript components like the OAuth handler.
Uses mkdir and mv for managing local skill modules and update directories.
[PERSISTENCE]: The skill provides instructions for users to store API credentials by exporting them in shell profile files.
Recommends adding BYBIT_API_KEY and BYBIT_API_SECRET to ~/.zshrc or ~/.bashrc.
Suggests path configurations in ~/.openclaw/.env for self-hosted environments.
[PROMPT_INJECTION]: The skill defines a safety hierarchy and defense mechanisms to handle external data safely.
Implements a "Rule Priority" where safety overrides other instructions.
Includes a specific "Prompt Injection Defense" section to prevent the agent from interpreting API response data (like P2P messages or order notes) as executable instructions.

bybit-trading