OpenMeta CLI

English | 简体中文

Turn open-source ambition into a repeatable contribution system.

OpenMeta CLI is a local-first autonomous contribution agent that helps developers discover worthwhile GitHub issues, prepare repository context, draft implementation plans, and compound contribution momentum over time.

Overview | Features | Quick Start | Commands | Roadmap | Contributing

Overview

Open source work usually breaks down long before code is written.

The real friction is often:

• deciding what to work on
• filtering noise from genuinely promising issues
• entering unfamiliar repositories quickly
• rebuilding context from scratch every time
• losing continuity between one contribution and the next

OpenMeta CLI is designed to reduce that friction and turn contribution work into a system instead of a one-off burst of energy.

It combines:

• issue discovery and ranking
• repository-aware workspace preparation
• patch and PR draft generation
• durable contribution artifacts
• local automation for recurring execution

Why OpenMeta

Built for contribution throughput

OpenMeta is not just an issue recommender. It is designed around the full path from opportunity discovery to contribution-ready artifacts.

Local-first by default

State, workspace preparation, memory, and generated artifacts stay local. The only remote services involved are the GitHub and LLM APIs you explicitly configure.

Designed to compound

Each run can improve the next one through repo memory, inbox state, and proof-of-work records. The system is meant to get more useful over time.

Features

Opportunity scoring

OpenMeta ranks GitHub issues against your technical profile using signals such as:

• stack fit
• freshness
• onboarding clarity
• merge potential
• repository impact

Repository-aware workspace preparation

Once an opportunity is selected, OpenMeta can:

• prepare a local workspace
• detect likely files
• extract useful snippets
• detect runnable validation commands
• carry repository memory into the next decision step

Draft generation

OpenMeta can generate:

• patch drafts
• PR drafts
• contribution dossiers
• inbox entries
• proof-of-work records

Durable contribution state

OpenMeta preserves:

• repo memory
• recent issue outcomes
• preferred file paths
• artifact history
• proof-of-work logs

Local automation

OpenMeta can install and manage a local daily automation workflow:

• launchd on macOS
• cron on Linux
• manual fallback on unsupported platforms

Workflow

flowchart TD
    A["Scout GitHub Issues"] --> B["Score Fit + Opportunity"]
    B --> C["Select Target"]
    C --> D["Prepare Local Workspace"]
    D --> E["Inspect Files + Checks"]
    E --> F["Draft Patch Strategy"]
    F --> G["Draft PR Narrative"]
    G --> H["Save Memory / Inbox / Proof of Work"]
    H --> I["Optional Local Automation"]

Loading

Architecture

Current project layout:

src/
  cli.ts              # entrypoint
  commands/           # command wiring
  contracts/          # structured agent contracts
  infra/              # config, logger, prompts, helpers
  orchestration/      # init, agent, config, automation flows
  services/           # github, llm, workspace, memory, scheduler
  types/              # shared config and domain types
test/                 # Bun test suite
bin/                  # built CLI output

Requirements

• Bun 1.0+
• Git
• GitHub Personal Access Token
• An LLM API key

Supported LLM options currently include:

• OpenAI
• MiniMax
• Kimi (Moonshot AI)
• GLM (Zhipu AI)
• Custom OpenAI-compatible endpoints

Installation

Install dependencies:

bun install

Build the CLI:

bun run build

Create a global openmeta command on Linux or macOS:

ln -sf "$(pwd)/bin/openmeta.js" /usr/local/bin/openmeta
chmod +x ./bin/openmeta.js
hash -r

Verify the installation:

openmeta --help

Quick Start

Initialize configuration:

openmeta init

Run the autonomous contribution workflow:

openmeta agent

Scout only:

openmeta scout --limit 10

Inspect durable artifacts:

openmeta inbox
openmeta pow

Review configuration:

openmeta config view

If you prefer not to create a global symlink, you can still run the built file directly:

./bin/openmeta.js init
./bin/openmeta.js agent --run-checks
./bin/openmeta.js automation status

Command Reference

Command	Description
openmeta init	Initialize OpenMeta CLI configuration
openmeta agent	Run the autonomous contribution workflow
openmeta agent --headless	Run unattended using saved automation defaults
openmeta agent --run-checks	Execute detected baseline validation commands
openmeta daily	Compatibility alias for agent
openmeta scout --limit <count>	Rank the highest-value contribution opportunities
openmeta inbox	Show drafted contribution opportunities
openmeta pow	Show proof-of-work history
openmeta runs	Show recent command runs, durations, and failure reasons
openmeta runs <id>	Inspect one recorded run
openmeta provider list	List saved LLM provider profiles
openmeta provider config	Configure a provider profile interactively
openmeta provider save <name>	Save current LLM settings as a reusable provider profile
openmeta provider add <name>	Add a provider profile from command-line values
openmeta provider use <name>	Switch the active LLM provider to a saved profile
openmeta provider remove <name>	Remove a saved provider profile
openmeta automation status	Show automation state
openmeta automation enable	Enable unattended daily automation using saved settings
openmeta automation disable	Disable unattended daily automation and remove the system scheduler
openmeta config view	Show current configuration
openmeta config set <key> <value>	Set a configuration value
openmeta config reset	Reset configuration to defaults

Example provider workflow:

openmeta provider config
openmeta provider save production
openmeta provider use production --validate

# optional: tune context budget and autonomous mode directly
openmeta config set llm.maxContextTokens 262144
openmeta config set automation.autonomousAgentEnabled true

Local Paths

OpenMeta maintains a clear local footprint:

• config: ~/.config/openmeta/config.json
• workspaces: ~/.openmeta/workspaces
• artifacts: ~/.openmeta/artifacts
• repo memory and proof-of-work state: stored in the local OpenMeta state area

Bounded Context Expansion

During openmeta agent, OpenMeta now retries implementation drafting with bounded repository context expansion when the model reports insufficient code context.

• Context expansion is enabled by default and only reads files inside the prepared workspace.
• Standard runs stop at three expansion rounds; autonomous mode can extend the loop to five rounds with a wider snippet budget.
• The loop never bypasses dirty-workspace, draft-only, validation, or generated-file safety gates.
• Editable context is capped by the active provider profile's maxContextTokens setting using deterministic compression; no extra summarization model call is required.

Security Model

OpenMeta is built around user-controlled execution:

• GitHub PAT and LLM API keys are stored with AES encryption
• no OpenMeta-hosted backend is required
• only explicitly configured GitHub and LLM services are contacted
• unattended automation is opt-in

Validation

Typical local verification commands:

bun x tsc --noEmit
bun test
bun run build

Roadmap

Planned or actively evolving areas include:

• stronger real-PR publishing flows
• richer repository memory and quick-win scanning
• deeper provider coverage for OpenAI-compatible ecosystems
• more robust unattended automation and recovery behavior
• improved artifact publishing and contribution traceability

Contributing

Contributions are welcome.

Good contribution areas include:

• improving issue ranking and repo memory quality
• expanding provider support
• strengthening workspace preparation and validation logic
• refining automation and artifact publishing flows
• improving documentation and onboarding

Suggested local workflow:

git checkout -b feat/your-change
bun install
bun x tsc --noEmit
bun test

License

MIT