Agent Orchestration
- 504 installs
- 3.9k repo stars
- Updated January 26, 2026
- parcadei/continuous-claude-v3
agent-orchestration is an agent skill that spawns implementation sub-agents to preserve main session context across long-running development for developers who orchestrate multi-file Claude Code workflows.
About
agent-orchestration is a parcadei/continuous-claude-v3 skill defining rules for when Claude should delegate implementation to sub-agents instead of burning main context on file reads and edits. The anti-pattern consumes 2000+ tokens in the main thread reading, understanding, and editing directly; the recommended pattern spawns an implementation agent that returns a roughly 200-token summary after completing multi-file work. A task table maps multi-file implementation, complex refactors, and broad codebase exploration to agent delegation while keeping lightweight queries in main context. Developers reach for agent-orchestration when running continuous-claude sessions that must stay coherent across many turns without exhausting the primary context window. The skill is user-invocable false, meaning it applies automatically as orchestration policy rather than a manual slash command, shaping agent spawning behavior throughout extended build sessions.
- Enables continuous Claude sessions that survive across multiple prompts and days
- Maintains project-wide context without repeated onboarding
- Orchestrates specialized agent roles and handoffs automatically
- Reduces token waste from re-explaining codebase repeatedly
- Supports complex multi-agent workflows in one workspace
Agent Orchestration by the numbers
- 504 all-time installs (skills.sh)
- +3 installs in the week ending Aug 4, 2026 (Skillselion tracking)
- Ranked #1,757 of 16,546 AI & Agent Building skills by installs in the Skillselion catalog
- Data as of Aug 5, 2026 (Skillselion catalog sync)
npx skills add https://github.com/parcadei/continuous-claude-v3 --skill agent-orchestrationAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 504 |
|---|---|
| repo stars | ★ 3.9k |
| Last updated | January 26, 2026 |
| Repository | parcadei/continuous-claude-v3 ↗ |
How do you preserve context in long Claude sessions?
Maintain persistent context and orchestrate multiple Claude instances across long-running development sessions.
Who is it for?
Developers running long Claude Code sessions with multi-file implementations who need automatic sub-agent delegation policies.
Skip if: Single-file edits or quick questions where direct main-context work is faster than agent spawn overhead.
When should I use this skill?
User requests multi-file implementation, large refactors, or extended coding sessions in continuous-claude-v3 workflows.
What you get
Delegated implementation via sub-agents with compact summaries and preserved main session context.
- Sub-agent implementation summaries
- Preserved main session context
By the numbers
- Main-context direct edits can consume 2000+ tokens per implementation cycle
- Sub-agent summaries return roughly 200 tokens to main context
Files
Agent Orchestration Rules
When the user asks to implement something, use implementation agents to preserve main context.
The Pattern
Wrong - burns context:
Main: Read files → Understand → Make edits → Report
(2000+ tokens consumed in main context)Right - preserves context:
Main: Spawn agent("implement X per plan")
↓
Agent: Reads files → Understands → Edits → Tests
↓
Main: Gets summary (~200 tokens)When to Use Agents
| Task Type | Use Agent? | Reason |
|---|---|---|
| Multi-file implementation | Yes | Agent handles complexity internally |
| Following a plan phase | Yes | Agent reads plan, implements |
| New feature with tests | Yes | Agent can run tests |
| Single-line fix | No | Faster to do directly |
| Quick config change | No | Overhead not worth it |
Key Insight
Agents read their own context. Don't read files in main chat just to understand what to pass to an agent - give them the task and they figure it out.
Example Prompt
Implement Phase 4: Outcome Marking Hook from the Artifact Index plan.
**Plan location:** thoughts/shared/plans/2025-12-24-artifact-index.md (search for "Phase 4")
**What to create:**
1. TypeScript hook
2. Shell wrapper
3. Python script
4. Register in settings.json
When done, provide a summary of files created and any issues.Trigger Words
When user says these, consider using an agent:
- "implement", "build", "create feature"
- "follow the plan", "do phase X"
- "use implementation agents"
Related skills
Setup Matt Pocock SkillsScaffold the per-repo configuration that Matt Pocock’s engineering agent skills rely on so they understand the issue tracker, triage labels, and domain documentation la495k196k
Lark AppsConnect Claude, Cursor or custom agents directly to Lark (Feishu) for messaging, document automation, approval workflows and enterprise data access.402k
CavemanSlash token usage by roughly 75% while keeping every technical detail intact when working with Claude Code, Cursor or similar agents.391k94.7k
Lark Skill MakerQuickly turn any Lark/Feishu OpenAPI call or multi-step workflow into a reusable agent skill with its own SKILL.md.386k16k
Running Claude Code Via Litellm CopilotRun Claude Code at a fraction of the cost by routing requests through LiteLLM to the GitHub Copilot Chat API.270k72
Codex PetGenerate a complete Codex Pet spritesheet and metadata from one reference image without needing an OpenAI key or Codex Pro.246k10FAQ
Why does agent-orchestration spawn sub-agents?
agent-orchestration spawns implementation sub-agents so multi-file reads and edits happen outside main context, returning a compact summary near 200 tokens instead of consuming 2000+ tokens in the primary session.
Is agent-orchestration manually invoked?
agent-orchestration sets user-invocable to false, applying automatically as orchestration policy during continuous-claude-v3 sessions rather than requiring a user-triggered slash command.