The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is designed to run a local Node.js script, scripts/taibu.mjs, which serves as the primary interface for 15 different divination tools. The agent is instructed to use this script for all calculations to ensure accuracy and reproducibility.
[DYNAMIC_EXECUTION]: The wrapper script scripts/taibu.mjs dynamically imports a local JavaScript bundle (scripts/vendor/taibu-core.bundle.mjs) containing the core calculation engine. This is a standard architectural approach for distributing local calculation logic within a skill.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided inputs such as birth dates, locations, and personal queries, creating a surface for potential indirect prompt injection.
- Ingestion points: User data is ingested through JSON parameters passed to the taibu.mjs script via command-line arguments or standard input.
- Boundary markers: The skill includes clear instructions in the interpretation/ documents that mandate the agent use tool output as the "only source of truth" and avoid adding unverified details, which serves as a logical constraint on the agent's behavior.
- Capability inventory: The local tools are limited to reading input and generating text/data for the agent; the provided scripts do not contain network or file-writing capabilities.
- Sanitization: The input is structured as JSON and undergoes parsing, which provides basic validation against malformed payloads.

taibu-divination