Agent-Skills.md

Agent Skills: Orchestrator Capacity Reference

Private CCB orchestrator skill for requesting, inspecting, using, and releasing dynamic loop capacity through `ccb loop capacity` only.

UncategorizedID: bfly123/claude_code_bridge/orchestrator-capacity

Repository

bfly123/claude_code_bridge
SeemSeamLicense: NOASSERTION
3,210310

Install this agent skill to your local

pnpm dlx add-skill https://github.com/SeemSeam/claude_codex_bridge/tree/HEAD/docs/plantree/plans/agentic-loop-workflow/drafts/agentroles.ccb_orchestrator/adapters/ccb/skills/orchestrator-capacity

Skill Files

Browse the full folder contents for orchestrator-capacity.

Download Skill

Loading file tree…

docs/plantree/plans/agentic-loop-workflow/drafts/agentroles.ccb_orchestrator/adapters/ccb/skills/orchestrator-capacity/SKILL.md

Skill Metadata

Name
orchestrator-capacity
Description
Historical reference for script-owned loop capacity requests. Do not use from provider sessions.

Orchestrator Capacity Reference

This reference is retained for rolepack history only. It is not projected as an active provider skill for agentroles.ccb_orchestrator.

Loop capacity is now runner-owned. The orchestrator may recommend the semantic need for capacity, but the supervisor/runner script performs any actual capacity ensure, status, release, routing, and cleanup.

Provider Boundary

  • Do not run CCB commands.
  • Do not inspect runtime capacity from the provider session.
  • Do not release agents, mutate task status, import artifacts, or route work.
  • Do not invent agent names, windows, panes, provider profiles, or topology.

Reply Shape

When capacity information is needed, reply with semantic evidence only:

  • requested profiles and counts;
  • task packet and verification refs;
  • blockers or missing evidence;
  • release-readiness recommendation based on supplied worker/reviewer evidence.

The runner owns command execution and treats provider replies as evidence only.