Power Oracle

Agents

How agents should discover, call, and interpret Power Oracle.

Power Oracle is designed for direct software use.

Discovery Path

Agents should start with:

  • https://api.workcapacity.io/openapi.json
  • GET /movements

The OpenAPI schema defines the request and response contract.

GET /movements provides supported movement names, required inputs, alternative inputs, overrides, and assumptions.

  1. Read the OpenAPI schema directly.
  2. Inspect GET /movements for supported movement metadata.
  3. Gather split-aware workout details.
  4. Build a canonical POST /compute-power payload.
  5. Handle an optional 402 Payment Required challenge if paid access is enabled.
  6. Interpret session metrics, split metrics, movement rollups, summary output, and notes.

Payload Construction Rules

  • Treat each meaningful work segment as its own split.
  • Use top-level duration_seconds for full session elapsed time.
  • Use split duration_seconds for active work time only.
  • Put recovery time into rest_seconds_after when it is known.
  • Do not send server-managed version fields.

Skill Distribution

There is already a local Power Oracle skill pattern in this repository for agent usage and maintenance.

A ClawHub skill is planned at https://clawhub.com/skills/power-oracle, but it is not live yet.

Until then, agents should use the OpenAPI schema directly rather than waiting for a packaged skill distribution.