mvschwarz/openrig-world

Dernière activité source enregistrée: 27 octobre 2026 à 16:45

Catalogue SkillsMP mis à jour: 29 octobre 2026 à 18:30

skills collectés: 29

Étoiles GitHub: 15
Forks GitHub: 4

Installer avec un assistant IA

Copiez ce prompt dans l’assistant IA que vous utilisez.

Follow https://skillsmp.com/skill-install/prompt.md to install Agent Skills from https://github.com/mvschwarz/openrig-world.

README du dépôt

Fichier source du README : README.md

Lire le README source

Afficher le README complet

# OpenRig project world A context pack that helps a coding agent work on [OpenRig](https://github.com/mvschwarz/openrig) itself. It covers what the project is for, how a contribution gets from an idea to a merged pull request, and how to prove that a change works. Install it once. After that, any agent working in an OpenRig checkout can load it with one command. > **This map is not the territory.** It is short on purpose, it is incomplete, and it will drift. > The OpenRig source, the `rig` binary you are running, and the maps inside the product repository > are the authorities. Where this pack disagrees with them, they are right. Please open an issue > or a pull request here when you find a disagreement. ## Who it is for - People who want to contribute to OpenRig, and the coding agents they work with. - Claude Code and Codex are supported directly: `rig context profile` composes a walk for either. Agents on other runtimes can still read every file with `rig context get`, or straight from this repository, because the files are plain Markdown. You do not need a multi-agent rig to use this pack. One agent in one terminal is enough. ## How it fits with the rest of OpenRig's context OpenRig gives agents context in layers. This pack is one of them. | Layer | What it answers | Where it comes from | |---|---|---| | System World | How to be an agent inside OpenRig: identity, peers, the queue, the command surface | Ships with OpenRig. On a default install it selects the `world-public` and `onboarding-width` packs; `rig context list` shows what your install has. | | **Project world (this pack)** | How to build OpenRig: its purpose, how contributions move, how to prove a change | This repository | | Code maps and the `developing-openrig` skill | Where code lives, what is risky, what to run | The product repository: `ARCHITECTURE.md`, `docs/as-built/arteries.md`, `docs/as-built/test-layers.md`, `.claude/skills/` and `.agents/skills/` | | Your work | The issue or pull request in front of you | GitHub | This pack points at the product repository's maps instead of copying them. A copy would drift, and the maps in your checkout describe the code you are actually changing. ## Install You need OpenRig installed, with its daemon running. The [product README](https://github.com/mvschwarz/openrig#install-and-first-run) covers requirements and [what OpenRig changes on your machine](https://github.com/mvschwarz/openrig#what-openrig-changes-on-your-machine). ```bash rig context add --git https://github.com/mvschwarz/openrig-world.git rig context list ``` `--git` clones this repository using your existing Git setup, finds `manifest.yaml` at its root, and installs the pack under the manifest's name, `openrig-world`. Only the files the manifest declares are served to agents; this README is for you. To check for and take later changes: ```bash rig context source inspect openrig-world rig context source update openrig-world ``` ## Load it Run one of these yourself, or ask your agent to run it. Each prints the composed text for the agent to read. | When | Command | |---|---| | Starting work on OpenRig | `rig context profile openrig-world --situation fresh` | | Reviewing a pull request | `rig context profile openrig-world --situation fresh --profile reviewer` | | After a compaction, `/clear` or restart | `rig context profile openrig-world --situation post-compaction` | | One file | `rig context get openrig-world/craft/proving-a-change.md` | | One section | `rig context get 'openrig-world/workflow/contributing-with-agents.md#after-a-context-reset'` | | Everything, in one bundle | `rig context get openrig-world` | `rig context profile` composes for Claude Code by default. Add `--runtime codex` for Codex, or set `OPENRIG_RUNTIME`. Add `--json` to see token estimates and content hashes for each piece. ## What is inside | File | What it covers | |---|---| | `identity/what-openrig-is-for.md` | What OpenRig is for, and how proposed changes are judged | | `workflow/contributing-with-agents.md` | The loop from install to merged pull request, done with agents | | `craft/proving-a-change.md` | How to show a change works: seams, checks that can fail, proof by effect | | `craft/tells-of-a-wrong-claim.md` | Moments when you are about to be wrong, and the cheapest check for each | | `manifest.yaml` | The pack manifest: files, atoms, and the `reviewer` profile | ## Rig catalog Reviewed listings, submissions, team diagrams and public status live in [openrig-registry](https://github.com/mvschwarz/openrig-registry). The workshop and factory-rsi bundle sources remain here under `rigs/`; their installation links do not change. `registry/workshop.yaml` is the only catalog file retained here, for OpenRig 0.6.6 operators. The copy in openrig-registry is canonical. Update this compatibility mirror with the exact same bytes in a paired, reviewed pull request when that pin changes, and merge both before announcing the new pin. See the [registry's compatibility instructions](https://github.com/mvschwarz/openrig-registry#workshop-compatibility-with-openrig-066). ## Naming a team Use the [topology naming reference](https://github.com/mvschwarz/openrig/blob/95412717f319c6557c970d870cc22fa976565837/docs/reference/topology-naming.md) when authoring rigs: a rig names the purpose, a pod names a domain, and a seat names its role. Starter grows from `dev.build` and `dev.review` to Workshop's lead and development pod, then Factory's independent review pair. Runtime presets keep those identities. The `factory-rsi` bundle remains an original workflow reference with its original addresses. It is distinct from the current built-in Factory and is not the naming template for a new team. Its bundle, workflow targets and pinned registry view remain together; old fixtures and run records keep the identities they describe. ## Contributing to this pack Pull requests are welcome, especially when you find something here that is stale or wrong. Keep pages short, point at sources instead of copying them, and describe commands only after you have run them. If you change `manifest.yaml` or a section heading that an atom addresses, commit, then check that your clone still loads under a scratch name: ```bash rig context add --git /path/to/your/clone --name openrig-world-test rig context profile openrig-world-test --situation fresh --json rig context profile openrig-world-test --situation post-compaction --json rig context profile openrig-world-test --situation fresh --profile reviewer --json rig context rm openrig-world-test ``` Each `profile` command should list its pieces with no error. ## License Apache-2.0, the same as the OpenRig product repository. See [LICENSE](LICENSE).

Chargement du README du dépôt…

Skills dans ce dépôt

classification en attente

skill

métier

description

mis à jour

queue-handoff

non classé

Use when ending a turn, finishing a slice, blocked on another agent's work, or escalating to a human — durable work handoff via queue items so the system keeps moving across compactions, missed messages, and interruptions. Covers the hot-potato…

Langue du texte source : anglais

2026-10-27

openrig-architect

non classé

Use when designing multi-agent topologies that run ON OpenRig — authoring RigSpec and AgentSpec files for new rigs, creating agent startup content (guidance / skills / culture), or diagnosing why a launched rig's agents aren't behaving as intended. NOT for…

Langue du texte source : anglais

2026-10-08

openrig-software-factory

non classé

Use when a user wants a continuing software team for a real repository, or has a first OpenRig team and needs a repeatable path for reviewed work and later tasks.

Langue du texte source : anglais

2026-10-08

using-the-tui-together

non classé

Use when you notice something the person would want to see (a seat waiting on them, a seat that has stalled or keeps failing, work that needs their decision), when they ask what the team is doing or where some work stands, or in their first session with this…

Langue du texte source : anglais

2026-10-07

cross-host-rig-commands

non classé

Use when addressing a registered remote OpenRig host, choosing its transport, or interpreting a cross-host result.

Langue du texte source : anglais

2026-10-06

messaging-the-human

non classé

Use when project policy calls for a human decision or update, a human delivery is pending or failed, or a reply must resume the right work.

Langue du texte source : anglais

2026-10-06

openrig-herdr

non classé

Use when opening OpenRig fleet terminals as a herdr wall — turning a rig, pod, mission, slice, or saved view into live interactive agent tiles via `rig terminal`, watching another rig read-only, or driving herdr on an agent's request ("open all my rigs + a…

Langue du texte source : anglais

2026-10-06

openrig-skills

non classé

Use when you're operating OpenRig and need the right skill or context for fleet recovery, seat handover, new-seat orientation, a watchdog wake, cross-host reach to an agent on another machine, rig packaging, an OpenRig upgrade, systematic debugging, queue…

Langue du texte source : anglais

2026-10-06

openrig-upgrade

non classé

Use when an agent is preparing or performing an OpenRig CLI/daemon upgrade and must preserve live seats, verify each mutation, or reconcile managed plugin and skill files without overwriting local work.

Langue du texte source : anglais

2026-10-06

rig-bundles-and-shareable-artifacts

non classé

Use when authoring or installing a rig bundle (packaged, shareable artifact that instantiates an opinionated OpenRig topology + workflow), reasoning about the bundle vs extension boundary, or auditing a bundle for portability. Covers the 4 failure modes that…

Langue du texte source : anglais

2026-10-06

rig-lifecycle

non classé

Use when reasoning about the rig lifecycle operations family (create / start / stop / resume / restore / snapshot / release / unclaim / destroy), reading or trusting `rig ps` / lifecycle projections after recovery, or designing proof for a lifecycle scenario.…

Langue du texte source : anglais

2026-10-06

topology-mutation-and-seat-management

non classé

Use when changing a rig while it is alive — `rig grow` / `rig expand` / `rig shrink` / `rig launch` / `rig remove` / `rig discover` / `rig bind` / `rig adopt` / `rig attach`. Covers the 4 failure modes (newly created seat lacks queue/startup/role; edges and…

Langue du texte source : anglais

2026-10-06

development-team

non classé

Use when a development pod begins or hands off implementation, QA, or design work.

Langue du texte source : anglais

2026-10-06

orchestration-team

non classé

Use when coordinating assignments, selected review boundaries, or blocked work across a rig.

Langue du texte source : anglais

2026-10-06

review-team

non classé

Use when assigned a review or when an authored review boundary is reached.

Langue du texte source : anglais

2026-10-06

agent-browser

non classé

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include…

Langue du texte source : anglais

2026-10-06

dogfood

non classé

Systematically explore and test a web application to find bugs, UX issues, and other problems. Use when asked to "dogfood", "QA", "exploratory test", "find issues", "bug hunt", "test this app/site/platform", or review the quality of a web application.…

Langue du texte source : anglais

2026-10-06

systematic-debugging

non classé

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes

Langue du texte source : anglais

2026-10-06

test-driven-development

non classé

Use when implementing any feature or bugfix, before writing implementation code

Langue du texte source : anglais

2026-10-06

verification-before-completion

non classé

Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always

Langue du texte source : anglais

2026-10-06

ponytail

non classé

Forces the laziest solution that actually works, simplest, shortest, most minimal. Channels a senior dev who has seen everything: question whether the task needs to exist at all (YAGNI), reach for the standard library before custom code, native platform…

Langue du texte source : anglais

2026-10-05

notes

non classé

Use when taking notes.

Langue du texte source : Indéterminée

2026-10-05

openrig-user

non classé

Use when a specific `rig` command, subcommand, or flag is already known and you need its exact syntax, JSON shape, defaults, or error meaning. NOT for natural capability discovery, open-ended how-do-I questions, or choosing which OpenRig move applies.

Langue du texte source : anglais

2026-10-08

openrig-skills

non classé

Langue du texte source : anglais

2026-10-06

applying-a-permission-policy

non classé

Use before launching a team during agent-guided setup, or when a user asks to configure OpenRig command permissions, reduce repeated native approval prompts, or apply a selected rig/seat permission policy.

Langue du texte source : anglais

2026-10-05

claude-compaction-restore

non classé

Use when a Claude Code session is about to compact, has just compacted, reached its context limit, resumed after /compact, or must rebuild its working picture of a long-running seat from its own restore map, its JSONL transcript and the files it touched.

Langue du texte source : anglais

2026-10-05

forming-an-openrig-mental-model

non classé

Use when the system around you does not make sense yet: you just booted into a seat and do not know how the pieces fit; someone said rig, pod, seat, fleet, topology, or slice and you are not certain what they mean here; you are unsure what kind of rig you are…

Langue du texte source : anglais

2026-10-05

seat-continuity-and-handover

non classé

Use when replacing a seat's occupant (rebuild/handover/swap), reasoning about stable-seat-identity vs fluid-occupant-identity, choosing an old-occupant disposition (retire/advise/shadow), or recording provenance for an occupant change. Two independent…

Langue du texte source : anglais

2026-10-05

session-compaction-and-restore

non classé

Use when reasoning about what survives compaction/context-loss/restart/seat-refresh, designing a high-fidelity restore packet, or distinguishing native runtime resume vs fork vs artifact-backed mental-model rebuild. Covers the 4 failure modes that prevent…

Langue du texte source : anglais

2026-10-05

Affichage de 29 skills collectés sur 29.