The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
Brigade
Track work, verify results, carry memory, and sync tools across coding agents. Local files, no daemon.
Install
git clone https://github.com/escoffier-labs/brigadeInstall command taken from the project's README or registry entry. Always review code before running it.
README excerpt
Read full README on GitHub →Brigade keeps coding-agent work trackable across tools, repos, and sessions: tasks, receipts, shared memory, and synced tools.
Install · Docs · Compare · Quickstart
Published v0.27.0 is the stable install (brigade-cli==0.27.0). It adds per-repo parallel-safe work waves, cross-repo ready-work campaigns, Memory Operations, an optional fleet hub, care-managed memory jobs, cloud seats with receipts, vault projection, run lineage, code-graph views, bounded recall, declared budgets, and a machine-readable control crosswalk. Main is now the 0.28 beta line.
What it does
- Ready work and claim safety. Brigade lists unblocked tasks and uses atomic claims so two agents cannot take the same item.
- Declared run limits, optional fleet hub, and registered cloud activity. Brigade enforces declared wall-clock and worker-dispatch ceilings, records named retry and reroute decisions, can report runs, repo claims, and external harness sessions through an operator-owned fleet hub, and reconciles registered cloud work against provider and GitHub state.
- Verification receipts. A check run through Brigade writes the command, the real exit code, and the Git state. Not every Brigade command writes a receipt.
- Owner-mediated handoffs and bounded recall. Handoffs are linted and routed by the memory owner. Safe targeted notes may auto-file, while ambiguous or risky notes wait for review. Session-start recall is capped, and vault project, search, and propose are bounded.
- Explicit dry-run-first harness projection. MCP servers, tools, and skills are previewed before Brigade writes them into a harness.
Learning in Brigade is outcome-based skill scoring, promotion, and rollback from captured verification receipts. It is not autonomous reflective memory learning.
Install
Install with pipx install brigade-cli or uv tool install brigade-cli, then follow the install guide, QUICKSTART.md, or give first 10 minutes and docs/agents-guide.md to your coding agent.
Stable channel: published v0.27.0. Preview channel: brigade update --channel beta for 0.28.0.devYYYYMMDD wheels. Details: update channels.
How it compares
Brigade overlaps several categories. Most adjacent projects specialize in one layer. Brigade connects those layers around trackable coding-agent work. Where a neighbor is stronger, that cell says so. Dated sources and longer notes: docs/comparison.md (2026-08-27).
Comparison key: Supported · Partial · Not documented
Work and orchestration
Similar plugs
Reverse engineer anything with agents, from app behavior down to native binaries.
Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype…
User-friendly AI Interface (Supports Ollama, OpenAI API, ...)
A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build & Hermes Agent. Only official website: ccswitch.io