A monorepo workshop for agent-development tooling: CLIs, linters, and playbooks built while shipping AI-agent coding workflows in practice. Two packages ship on npm; the rest are working tools used inside this workshop and its sibling repos.
Installer for an orchestrator-led agent workflow: a .ai/ directory for run state, a marker-fenced policy section in AGENTS.md, and subagent definitions with preselected models for Claude Code, OpenAI Codex, and opencode.
npx orchestrator-workflow initSee packages/orchestrator-workflow for the full install and usage guide.
Validates knowledge bundles against the Open Knowledge Format (OKF) v0.1 spec: frontmatter shape, reserved files, link resolution, absolute-link warnings, sources shape.
npx okf-kit check path/to/bundleSee packages/okf-kit for the full install and usage guide.
slop-detector is the AI-slop linter for PRs: it catches leaked tool-call XML, doubled ## Summary headings, hedging openers, marketing adjectives, JSDoc on trivial getters, and other agent-generated tells across six rule packs (agent-tics, prose-slop, comment-slop, code-slop, ui-slop, placement-slop; some packs are opt-in). It runs in pre-commit, in CI as a status check, or ad-hoc against a path.
Not yet published to npm (the bare slop-detector name belongs to an unrelated third-party package), so it runs from a local build:
cd packages/slop-detector && npm install && npm run build && cd ../..
node packages/slop-detector/dist/cli.js check examples/slop-sample.md --explainFull pack reference, sample output, the scan pipeline, and the rationale (including a data point from running it against 20 recently merged PRs): packages/slop-detector/README.md.
| Package | What it does | npm |
|---|---|---|
| orchestrator-workflow | Installer for an orchestrator-led agent workflow: .ai/ run state, an AGENTS.md policy section, and subagent definitions with preselected models for Claude Code, Codex, and opencode. |
published |
| okf-kit | CLI that validates OKF v0.1 knowledge bundles: frontmatter shape, reserved files, link resolution, absolute-link warnings, sources shape. |
published |
| slop-detector | AI-slop linter for PRs: leaked tool-call XML, doubled Summary headings, hedging openers, marketing adjectives, and more across six rule packs. | not published (name taken; run from a local build) |
| agent-primitives | Agent-first CLI: bounded JSON envelopes, a mutation-probe runner, a verify runner, and a PATH doctor, plus an init command that installs its own skill document into a harness's skill directory. |
published |
| agent-dev-kit | CLI scaffolding for AI agent projects: file layout, hooks, entrypoints. | not published |
| friction-log | Capture, query, and infer agent-workflow frictions. SQLite-backed, sink-pluggable, zero-config default. | not published |
| github-api-tool | TypeScript CLI for GitHub API operations (issues, PRs, commits, standup digests), JSON output for agents calling via exec. |
private |
| git-batch-cli | Run safe batch git operations across all repos under a folder: sync, status, dirty checks, fetch, with --strict for automation. |
not published |
| mcp-token-audit | Ranks tool calls in Claude Code transcripts by approximate token cost per tool name, with an mcp__* share of the total. |
not published |
| agent-engineering-playbook | Guide for building production-ready AI agent systems. | doc package |
| agentic-coding-playbook | Practical playbook for teams using AI agents in coding. | doc package |
agent-dx is a folder of independent packages, not an npm workspaces / pnpm / lerna monorepo. There is no root package.json, no workspace manifest, and no shared root node_modules. Each package under packages/ carries its own package.json, install, build, test, and version, so the install pattern for any local-build package is the same one shown above for slop-detector:
cd packages/<name> && npm install && npm run buildIf you only care about one package, work in its directory; nothing at the root needs to be set up first.
Experimental: each package has its own version, README, and CI. APIs may evolve at minor-version bumps. orchestrator-workflow and okf-kit are published to npm with tagged releases. slop-detector is deliberately unpublished (the bare name on npm belongs to an unrelated package) and ships an MCP server alongside the CLI. agent-dev-kit, friction-log, git-batch-cli, and mcp-token-audit are functional CLIs, not yet on npm. github-api-tool is marked private in its own package.json. agent-engineering-playbook and agentic-coding-playbook are documentation packages, not code.
slop-detector and the workshop around it contribute the authoring-side tooling to the Project OS human-agent dev lifecycle. It sits alongside:
- agent-planforge plans
- agent-tasks coordinates
- agent-grounding verifies (evidence ledger, claim gate, hypothesis tracker)
- agent-preflight gates pushes
- harness declares + enforces the policy boundary that calls into all of the above
scaffoldkit and agent-planforge are standalone tools used by project-forge.