By pleri
Olam — agentic thinking platform. Spawn isolated development worlds, dispatch tasks to coding agents inside them, and capture the thought graphs they produce.
Commit changes, push the branch, and open a PR using the repo's template. Use when the user says "ship this", "open a PR", "commit and push", "make a PR for this", or when work is complete and ready for review. Analyses the diff, generates conventional commit messages (multi-commit detection), and uses the repo's `.github/PULL_REQUEST_TEMPLATE.md`.
Connect Claude Code to a remote admin-hosted Olam deployment via the MCP `http` transport (Worker URL + CF Access SSO). Use when the user has a Worker URL from an admin OR says "connect Claude to olam", "add olam MCP via URL", "set up admin-hosted olam", "wire up olam mcp endpoint". NOT for the standard shuk-marketplace install (that auto-registers a local stdio MCP via `.mcp.json`); NOT for shelling into a running world's container (see `olam-enter`).
Create a new isolated development world (local Docker or Cloudflare Sandbox). Use when the user says "create a world", "new world", "spin up an environment", "start a fresh sandbox", "I need an isolated workspace for a task", or wants to dispatch an agent into a sandboxed branch. Selects the compute backend from the project's Olam config; both modes flow through one MCP tool.
Destroy an Olam world and clean up all resources (containers, worktrees, services, R2 artifacts). Use when the user says "destroy world", "tear down", "clean up world", "kill that world", "remove environment", or wants to free resources held by a world they no longer need. Confirms current state before destroying.
Dispatch a task prompt to a running Olam world's Claude Code agent. Use when the user says "dispatch", "send task to world", "tell world X to do Y", "run this prompt in world X", or wants the agent inside a specific world to pick up new work. Operates on already-running worlds; for creating a new world use `olam-create`.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
@pleri/olam-cliCLI for Olam — the agentic-thinking platform. Spawn isolated development worlds, dispatch tasks to coding agents inside them, capture thought graphs they produce, ship PRs from inside worlds, and orchestrate parallel sub-tasks ("lanes") under one world.
This is the published artifact repository. Source lives in pleri/olam. Filing issues against this repo is fine for install/distribution problems; functionality issues belong in the source repo.
claude plugin marketplace add pleri/shuk
claude plugin install olam@shuk
Restart Claude Code. The MCP server boots on first /olam:* slash command. See plugin/README.md for plugin install paths (shuk, olam mcp install, upgrade, offline git-clone fallback) + the dual-mode .mcp.json rationale.
curl -fsSL https://olam.bar.dev/install | sh
The installer:
npm prefix and verifies it's writable (offers a user-local fallback if not)npm against GitHub Packages using your gh CLI token (refreshes scope if missing read:packages)npm install -g @pleri/olam-cli@stableolam setup to finish per-machine bootstrap# 1. Authenticate npm against GitHub Packages
npm login --scope=@pleri --registry=https://npm.pkg.github.com
# 2. Install
npm install -g @pleri/olam-cli@stable
# 3. Bootstrap
olam setup
You'll need a GitHub PAT with read:packages scope (and access to the pleri org).
Olam organizes capability into a small number of command groups. Every group has --help.
olam create my-task --task "implement feature X" # spawn a new isolated world
olam list # show running worlds
olam status [<world-id>] # machine status or per-world details
olam enter <world-id> # open a terminal inside a world
olam destroy <world-id> # tear down a world and clean up
olam clean # reclaim stale worlds + orphan resources
olam observe <world-id> # stream the agent's reasoning (live)
olam dispatch <world-id> "next prompt" # send a follow-up prompt to a running world
olam crystallize <world-id> # save the world's thought graph to the Pleri Plane
olam lanes list <world-id>
olam lanes create <world-id> <lane-name>
olam lanes dispatch <world-id> <lane-name> "prompt"
olam lanes destroy <world-id> <lane-name>
olam lanes merge <world-id> <lane-name> # merge a lane's branch back into the world
olam pr list <world-id> # PR-gate requests from the world's agent
olam pr show <world-id> <pr-id>
olam pr track <pr-id> # follow status of a PR you've filed
olam pr approve <pr-id>
olam pr reject <pr-id> --reason "..."
olam workspace list
olam workspace add <name> --path <dir>
olam workspace show <name>
olam workspace remove <name>
olam repos list # repos visible in the active workspace
olam repos add <git-url>
olam repos remove <name>
olam repos update # pull all
olam kg build [<world-id>] # build the graphify-backed graph
olam kg status [<world-id>] # build progress + freshness
olam kg watch [<world-id>] # re-build on file change
olam kg classify <world-id> --rules <file> # apply classification rules
olam kg doctor [<world-id>] # diagnose graph health
Agents inside worlds query the graph via the registered MCP server; prefer it over raw grep for symbol-shaped questions (see docs/decisions/014-kg-first-search.md in the source repo).
olam runbooks list
olam runbooks show <name>
olam runbooks add <name> --file <markdown>
olam runbooks apply <name> <world-id>
olam runbooks remove <name>
olam auth up # bring the auth-service vault up
olam auth down # take it down
olam auth status
olam auth login <provider> # claude, github, npm
olam auth logout <provider>
olam auth refresh <provider>
olam auth list
olam auth disable <id>
olam auth enable <id>
olam auth remove <id>
npx claudepluginhub pleri/olam-cliMulti-project workspace orchestration — manage projects, git operations, encrypted secrets, and LLM wikis (Karpathy pattern) across a unified workspace.
Permanent coding companion for Claude Code — survives any update. MCP-based terminal pet with ASCII art, stats, reactions, and personality.
Access thousands of AI prompts and skills directly in your AI coding assistant. Search prompts, discover skills, save your own, and improve prompts with AI.
Reliable automation, in-depth debugging, and performance analysis in Chrome using Chrome DevTools and Puppeteer
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Intelligent draw.io diagramming plugin with AI-powered diagram generation, multi-platform embedding (GitHub, Confluence, Azure DevOps, Notion, Teams, Harness), conditional formatting, live data binding, and MCP server integration for programmatic diagram creation and management.
Persistent file-based planning for AI coding agents. Crash-proof markdown plans (task_plan.md, findings.md, progress.md) that survive context loss and /clear, with an opt-in completion gate and multi-agent shared state. Manus-style. Works with Claude Code, Codex CLI, Cursor, Kiro, OpenCode and 60+ agents via the SKILL.md standard. Includes Arabic, German, Spanish, and Chinese (Simplified and Traditional).