agentleFS
Sign inStart free
Starlight-IntelligenceStarlight-Intelligence/Starlight-Intelligence-System/skills/orchestration/starlight-teammates/SKILL.md

starlight-teammates

Skill10 starsChanged 2 months agoMITView on GitHub

Coordinate autonomous sovereign teammates across the Starlight Intelligence System. Routes tasks across dynamic frontier models (Sonnet 5.5, Gemini 4 Argon, Opus 5.5, GPT-6.1 Sol), enforces fail-closed governance invariants, integrates Refero MCP and BrowserOS Neo, and anchors state to the 6-vault memory substrate.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    When coordinating high-autonomy background work across sandbox environments (Firecracker / E2B / BrowserOS Neo)

    line 16
  2. Distinctive:

    Before executing any tool or state change, check governance gates:

    line 46
  3. Distinctive:

    If new architectural patterns or decisions emerged, commit to Technical or Strategic Vaults with SIP attestation footer.

    line 60

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Starlight Teammates Orchestration

"Autonomous execution with sovereign memory and fail-closed governance."

When This Skill Activates

  • Task involves autonomous digital teammates or workers ("teammate", "teammates", "queen teammate", "search manager", "github officer", "community lead", "computer teams", "dispatch teammate")
  • User runs starlight teammates CLI commands or interacts with /teammates cockpit route
  • Invocations of MCP tools: sis_teammates_list, sis_teammates_dispatch, sis_teammates_approve
  • When coordinating high-autonomy background work across sandbox environments (Firecracker / E2B / BrowserOS Neo)

The Sovereign Teammates Fleet

TeammateRoleDomainDefault TierFrontier ModelsActive Vaults
Starlight QueenEcosystem Operator & CIOorchestrationStandardClaude Sonnet 5.5, Opus 5.5, Gemini 4 ArgonOperational, Wisdom, Strategic, Technical
Search ManagerIntelligence & Lab AnalystintelligenceLightGemini 4 Argon, GPT-6.1 Sol, GLM 5.3Strategic, Wisdom
GitHub OfficerCode Quality & Sentinel GuardiansafetyStandardClaude Sonnet 5.5, GPT-6.1 SolTechnical, Operational
Community LeadMember Engagement & SynthesizerrelationalLightGemini 4 Argon, GLM 5.3Creative, Wisdom
Teammate BuilderSwarm Node & Skill CompilerfoundryFrontierClaude Opus 5.5, GPT-6 Astra, Sonnet 5.5Technical, Wisdom

Execution Procedures

Procedure 1: Fleet Inventory & Task Intake

  1. Inspect available sovereign teammates via teammatesEngine.getTeammates() or sis_teammates_list.
  2. Map incoming request intent to the appropriate teammate specialist.
  3. Pass input through The Veil (SanitizationGateway) to scrub credentials, tokens, and PII before processing.

Procedure 2: Dynamic Multi-Model Routing

Evaluate intent complexity and route dynamically to model tiers:

  • Light Tier (Gemini 4 Argon, GPT-6.1 Sol, GLM 5.3): Fast web search synthesis, linting, status checks, JSON extraction, background heartbeat.
  • Standard Tier (Claude Sonnet 5.5): Primary daily driver, full-stack code implementation, PR reviews, multi-tool workflows.
  • Frontier Tier (Claude Opus 5.5, GPT-6 Astra): Deep systems architecture, mathematical proofs, model council consensus, multi-repo refactoring.

Procedure 3: Fail-Closed Governance Invariants (Non-Waivable)

Before executing any tool or state change, check governance gates:

  1. Financial Transactions: Any prompt or action touching credit cards, payments, invoices, or subscriptions MUST HALT immediately. Status changes to waiting_approval. Approval card emitted to cockpit.
  2. Production Deployments / Main Merges: Any action attempting direct push or merge to main branch MUST HALT awaiting human operator signature.
  3. Execution Freeze: Teammate halts execution cleanly until the operator resolves the gate via sis_teammates_approve or the UI approval card.

Procedure 4: Tool & Sandbox Orchestration

Teammates execute actions within isolated environments:

  • Refero MCP: Query live UI/UX design patterns from leading apps (refero_search_screens, refero_search_flows).
  • BrowserOS Neo: Autonomous web tasks in an authenticated operator browser profile (browseros-neo).
  • MicroVM Sandboxes: Isolate bash execution, file writes, and build pipelines in ephemeral Firecracker / E2B microVMs.

Procedure 5: Sovereign Memory Hydration & Settlement

  1. Prior to task execution, hydrate teammate context from relevant vaults (Strategic, Technical, Creative, Operational, Wisdom, Horizon).
  2. Upon task completion, record execution trace, models utilized, duration, and output receipt to the Operational Vault via sis_append_entry.
  3. If new architectural patterns or decisions emerged, commit to Technical or Strategic Vaults with SIP attestation footer.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free