| status | ACTIVE |
|---|---|
| version | 4.89 |
| owner_skill | ls-docs-create-move |
This is the public documentation map for LocalSetup. Start here when you want the install path, platform behavior, shipped skills, workflow model, or release/verification rules.
- Current version:
4.89.10 - Supported platforms:
codex, claude-code, cursor, kilo, opencode, openclaw, github-copilot-cli, github-copilot-vscode, cline-cli, cline-vscode, amp-cli, goose-cli, pi-cli, hermes-agent, qwen-code-cli, kimi-cli, factory-droid, antigravity-app, gemini-cli, omp-cli - Shipped skills:
115 - Workflow packages:
21 - Source:
ls/docs/_generated/facts.json
| Page | What it answers |
|---|---|
| Why LocalSetup | The operating-layer narrative, engineering background, and practical service philosophy |
| Licensing | Noncommercial use, commercial inquiries, and retained license boundaries |
| Project README | Why LocalSetup exists and why people should use it. |
| Quickstart | How to install, select platforms, verify, and update. |
| Command reference | Copy-paste installer, managed CLI, validation, and maintainer commands. |
| Common configuration | Shared schema, presets, precedence, migration and protected configuration boundaries. |
| Adaptive test safety | Memory headroom, debounced worker queues, cleanup and reusable engineering evidence. |
| LSCli operations | Explicit profiles, offline runtime/command setup, protected coding, sessions, recovery, branches, compaction and tool-free completion. |
| LSCli runtime contracts | Grants, disclosure, sandbox resources, broker protocols, durable evidence and completion schemas. |
| LSCli candidate qualification | Historical installed-artifact scenarios and their host/provider limits, separate from final published-release acceptance. |
| SDK source ownership | Private SDK payload, upstream provenance, dependency locks and artifact/SBOM verification. |
| Client integration metadata | Distinguish lifecycle, installation guidance, catalog support, and qualification evidence. |
| Deterministic client state | Resolve client state, allocate private artifacts, and verify restart bindings. |
| Features | Full capability catalog grouped by practical use. |
| Shipped skills catalog | All shipped skills with IDs, versions, and descriptions. |
| Bootstrap packs | Reusable Codex-first bootstrap prompts, pack metadata, audit boundaries, and future adapter entry points. |
| Workflow packages | How workflow packages differ from skills, install, validate, and generate docs. |
| Workflow package standard | Rules for first-class workflow packages and workflow.yaml. |
| Platform registry | Canonical platform IDs, paths, and adapter rules. |
| Adapter ownership | Shared adapter-directory ownership rules for install, repair, verify, detach, rollback, and migration planning. |
| Product positioning, voice, and branding | Plain-language product story, display names, technical identifiers, and visual identity. |
| Multi-platform install | Detailed install behavior and options. |
| Harness automation | Opt-in heartbeat activation, typed LSCli profiles, reserved actions/controller accounting, runtime artifacts, cron gating and command-policy boundaries. |
Read the current release guide for LocalSetup 4.89.10, including compatibility, updating, and verification. Find downloads in the latest published release.
The 4.4.0 guide covers the earlier context and package consolidation.
LocalSetup installs both capability skills and workflow packages into the managed package library. Keep WORKFLOW_PACKAGES.md as the canonical reference for:
- source layout and runtime shape
- installer behavior and dependency pull-in
- validation and generated workflow docs
| Page | What it covers |
|---|---|
| Workflow registry | Named workflows, aliases, impact expectations, and when to use them. |
| Workflow quick reference | Agent-facing workflow shortcuts and composite pipelines. |
| Decision tree workflow | Reverse-prompt planning: one question, options, preferred choice, rationale. |
| PRD schema and external agent guide | Spec format, outcome template, and external-agent handoff fields. |
| Git traceability | How PRDs, specs, outcomes, and commits stay connected. |
| Tmux ops managed workflow | Human and agent guide for managed tmux sessions, sudo probe, run IDs, logs, status, and cancellation. |
| Herdr operations | Master skill, owned pane/task lifecycle, remote setup, and reviewed provenance tracking. |
| Tmux ops remote guide | How to run human-visible ops when tmux is on another host. |
| Page | What it covers |
|---|---|
| Skills and rules | How always-loaded context, skills, and adapters fit together. |
| Agent Skills compliance | How LocalSetup implements the Agent Skills spec. |
| Skill importing | Import skills from URLs or local paths with validation and screening. |
| Skill discovery | Public skill registry/index workflow and recommendations. |
| Skill interoperability | Independent custom skills, provenance contracts, and portability between hosts. |
| Optional storage skills | Standalone Backblaze and Garage setup, execution, recovery, and verification boundaries. |
| Skill normalization | How imported or in-tree skills are cleaned up and standardized. |
| Task skill matching | How agents choose the right skill for a task. |
Registered capability highlights:
ls-nodejs-nextjs-development: Node.js/Next.js/React runbook for package-manager, build, migration, debugging, testing, security, deployment, and current-version verification.ls-github-starredrepos-sync: GitHub starred repository archive workflow for authenticated context checks, dry-run sync, repo scouting, metadata snapshots, and guarded publish flows.ls-shadcn-ui-components: shadcn/ui component workflow for setup, components, CLI/MCP, registry, theming, forms, aliases, Radix/Base UI, updates, and troubleshooting.ls-typescript-code-quality: TypeScript/TSX code quality, tsconfig, typed ESLint or Biome config, Node TypeScript scripts, and TypeScript-heavy framework code.ls-ui-browser-debugging: UI review and browser-driven debugging workflow for Chrome DevTools MCP, Playwright MCP/CLI, browser ownership, evidence capture, minimal fixes, and durable UI regression tests.ls-requesting-code-review: Use when requesting code review before merge or after substantial changes; provide focused requirements, diff range, and severity-calibrated review instructions.ls-omniroute-skills-update: OmniRoute update reporting for upstream skill discovery, LocalSetup comparison, provenance metadata, and report-first import/update/removal planning.
Agent Q is the bidirectional handoff layer for PRD/spec exchange over file_drop or mail with sealed payloads, registry checks, and ledgered outcomes.
| Page | What it covers |
|---|---|
| Shared OpenPGP runtime | Certificate, envelope, persistent authority, recovery and verified consumer contracts. |
| Agent-to-agent protocol | Transport principles, flows, pre-ship checks, and PRD field mapping. |
| Agent Q scenarios | Same-machine, multi-repo, local/remote, mail, and file_drop scenarios. |
| Agent Q build spec | Implementation order and backlog for the transport client. |
| Agent Q pattern | Queue layout and PRD file movement model. |
| Page | What it covers |
|---|---|
| Versioning | VERSION source of truth, Conventional Commits, and release sync. |
| Documentation alignment summary | Generated inventory, truth map, asset manifest, and docs-alignment audit results. |
| Harness automation | Explicit activation rules for autonomous harness capabilities. |
| Repository maintenance | GitHub rulesets, required checks, labels, triage, security settings, and release-maintenance gates. |
| Document lifecycle | ACTIVE, DRAFT, PROPOSAL, and deprecation rules. |
| Repo and data separation | What belongs in source vs. local/generated state. |
| Tooling policy | Python-first tooling expectations and public docs constraints. |
| Python architecture standard | Python architecture: new and substantially refactored Python tooling follows ls/docs/PYTHON_ARCHITECTURE_STANDARD.md; keep entrypoints thin, package responsibilities explicit, and existing debt baseline-managed. |