Skip to main content
Glama

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
SQUAD_MCP_WORKER_ENABLEDNoEnable background worker ACA Job for long runs.false
SQUAD_MCP_RUN_STATE_BACKENDNoBackend for run state storage: 'file' (local) or 'table' (Azure Table Storage).file
SQUAD_MCP_ENABLE_RENDER_PPTXNoOpt-in to enable the squad_render_pptx tool for PPTX rendering.false
SQUAD_MCP_REMOTE_PIPELINE_ENABLEDNoEnable the async advisory pipeline (squad_run, squad_status). Defaults to false.false

Instructions

Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.

This server publishes no instructions, or was last inspected before Glama recorded them.

Capabilities

Features and capabilities supported by this server

Protocol revision2025-11-25

CapabilityDetails
tools
{}

Tools

Functions exposed to the LLM to take actions

NameDescription
squad_researchA

Run the squad's research stage on a question or codebase. Routes to the Squad Researcher (squad researcher role) at the auto, parallel-eligible tier to investigate, explore, and gather evidence before any plan or change. Delegated execution: returns the Squad Coordinator persona, the matched routing row, and a framed dispatch request; the calling host runs the subagent loop. Use for "research", "investigate", "explore", or "find out" requests.

squad_planA

Run the squad's planning stage. Routes to the Squad Lead (squad lead role) at the confirm tier to break a task down, sequence the work, and produce an implementation-ready plan. Planning is non-parallel and lands no change without confirmation. Delegated execution: returns the Coordinator persona, matched routing, and a framed dispatch request for the host to run. Use for "plan", "break down", "sequence", or "design plan" requests.

squad_reviewA

Run the squad's review stage. Routes to the Squad Reviewer (squad tester role) at the auto tier for quality, correctness, and standards review. When the request asks for a pre-implementation go/no-go, crosses two or more council domains, or raises any responsible-AI concern, it engages a task-fit council at the confirm tier: only the seats whose lens the work touches (architect, security, cost-manager, product-owner, rai), with every lens left out recorded in the verdict. Delegated execution: returns the Coordinator persona, matched routing, and a framed dispatch request. Use for "review", "validate", "check quality", "council", or "go/no-go" requests.

squad_architectA

Run the squad's architecture stage. Routes to the System Architecture Reviewer (squad architect role) at the auto, parallel-eligible tier to review system design, component boundaries, and design tradeoffs (the Squad Azure Architect authors Azure HLD/LLD as the alternate). Delegated execution: returns the Coordinator persona, matched routing, and a framed dispatch request. Use for "architecture", "system design", or "components" requests.

squad_runA

Run the full squad pipeline for any request. Hands the request to the Squad Coordinator, which classifies it against the routing table and dispatches the matched roles through Research -> Plan -> Implement -> Review, carrying the council Implementation Gate, the autopilot/autonomous Human Gates, and the notification approval contract unchanged. This is the catch-all entry point and runs at the confirm tier with gates. Set mode=autonomous for the bounded validator loop or mode=autopilot for the full gated pipeline. Delegated execution: returns the Coordinator persona, matched routing, and a framed dispatch request for the host to drive. Use when no narrower squad_* tool fits, or for end-to-end work.

squad_federateA

Run the squad federation meta layer for a repository that hosts several named sub-squads (for example a product sub-squad for the business team and an azure sub-squad for the architects). Hands the request to the Squad Federation Coordinator, which reads the federation registry (federation.md) and meta-routing (meta-routing.md), selects the target sub-squad(s) — or honors an explicit squad=<name> — and runs each sub-squad's normal per-turn protocol scoped to .copilot-tracking/squad/members/<name>/. Set init to build a federation (propose → confirm → create) on a fresh project, or promote to adopt an existing single squad (a top-level team.md) into a federation as its first sub-squad, relocating its state intact. This is the catch-all federation entry point and runs at the confirm tier with gates. Delegated execution: returns the Federation Coordinator persona, the federation routing decision, and a framed dispatch request for the host to drive. Use for cross-sub-squad requests, an explicit sub-squad target, or federation setup.

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription

No resources

TDQS

A4.2/5.0

Scored across 6 tools

Disambiguation4/5

Tools are distinct by stage or entry point: research, plan, review, architect, federate, and run. The catch-all squad_run could conceptually overlap with stage-specific tools, but its description explicitly directs users to narrower tools when applicable, so ambiguity is minimal.

Naming Consistency5/5

All tools share the squad_ prefix followed by a single descriptive term (review, run, architect, etc.), yielding a predictable and consistent pattern. There is no mixing of camelCase or inconsistent verb/noun styles.

Tool Count5/5

6 tools is well-scoped for a squad orchestration server covering the main entry points and a federation meta layer. Each tool corresponds to a distinct stage or catch-all, with no redundancy.

Completeness3/5

The squad lifecycle includes Research, Plan, Implement, and Review, but there is no dedicated tool for the Implement stage. While squad_run covers the full pipeline including Implement, agents cannot directly trigger only implementation, leaving a notable gap.

Maintenance

ActivityActive
ResponsivenessUnresponsive