Rush
codexHand a task to a cloud agent that works in its own cloud computer and returns a pull request, a report, files or an answer; start, wait for, reply to and stop tasks through a hosted MCP server with OAuth sign-in.
6 skills · 1 MCP server
GitHub Copilot CLI agentOW harness with standard, POC, batch, review, and Accessibility workflows.
Plugin agentow-copilot from marketplace agentOW in kaixun96/dev.AgentOW. Commands read from the repo's manifest — 3 days ago.
This plugin runs code on your machine — it starts local MCP servers. The exact commands are listed below; read them before installing.
agentow-a11yFix an odsp-web Accessibility bug through a lightweight evidence-first pipeline. Codespaces skip host-only Windows AT tests unless an external evidence bridge is available; Windows hosts install safe scriptable prerequisites and run supported tests directly. After bounded evidence routes are exhaust
agentowTake an odsp-web feature or bug description to a draft PR on Azure DevOps. Use when the user wants to implement a feature, fix a bug, or make a change in the odsp-web monorepo. Drives the full pipeline: research → plan → implement → verify → fix loop → review → PR. Triggers on: implement, fix, add,
dethemeUse when implementing or fixing theming for a SharePoint surface such as app-chrome settings, a flyout, full page, dialog, callout, panel, drawer, or property pane. Classifies the surface, applies NeutralThemeProvider and only the required custom style hooks, handles v8 components and redundant Flue
ow-a11y-host-setupPrepare a Windows evaluator host for agentOW Accessibility testing. Installs scriptable dependencies, stages and launches the signed VB-CABLE driver installer, opens Voice Access first-run setup, and writes a capability report. Use for ow-a11y-host-setup, configure A11Y host, install VB-CABLE, confi
ow-batchRun multiple odsp-web agentOW tasks autonomously and serially in the current Copilot CLI session. Use when the user provides a list of features/bugs and wants each task to produce its own draft PR without stopping the whole batch on individual failures. Triggers on: ow-batch, batch agentow, run thes
ow-context-feedbackResume an agentOW run after user feedback and propose an evidence-grounded update to its linked context library. Triggers on: update context from feedback, agentOW feedback, revise context after PR feedback, remember what we learned from this run.
ow-doctorDiagnose and repair agentOW prerequisite problems. Use for ow-doctor, doctor, troubleshooting, broken FIC Playwright, missing MCP, broken Figma MCP, or bootstrap failures.
ow-initInitialize agentOW prerequisites once. Use for ow-init, initialize agentOW, first-time setup, install MCP prerequisites, or prepare agentOW.
ow-ref-external-toolsUse when agentOW needs a GUID, user alias, timestamp, killswitch blueprint, Bluebird search, ADO work item or PR data, ODSP wiki content, Microsoft Learn, or merge-conflict guidance.
ow-ref-monorepoUse for Rush, Heft, pnpm, rigs, project tags, lockfiles, package outputs, build cache, or ODSP-Web monorepo structure.
ow-ref-replace-componentUse for ODSP-Web UI component migrations, v8-to-SPDS replacements, Panel/Drawer conversion, stable-bundle adoption, or ReplaceComponent work.
ow-reviewRun the agentOW review gate on its own in Copilot CLI, outside the full pipeline. Use for ow-review, review this PR, review PR 1234567, re-review before publishing, review my branch, or standalone code review of odsp-web. Reviews an existing Azure DevOps PR by ID, or the current branch when no PR is
ow-share-insightsPreview and, only with explicit current-run user consent, prepare or send an anonymized AgentOW run-insights email.
vercel-react-best-practicesReact and Next.js performance optimization guidelines from Vercel Engineering. This skill should be used when writing, reviewing, or refactoring React/Next.js code to ensure optimal performance patterns. Triggers on tasks involving React components, Next.js pages, data fetching, bundle optimization,
a11y-evaluatorValidate Accessibility reproduction and post-fix evidence produced by an approved external evaluator or Windows-host evidence producer. This agent never controls assistive technology and never fixes code. It verifies scenario identity, evidence completeness, artifact hashes, and the strict reproduce
context-maintainerPropose evidence-grounded updates to the context library linked to an agentOW run. It writes an immutable candidate artifact; the main workflow applies it according to the linked context library's non-blocking update policy. - view - grep - glob - shell
evaluatorProactively dispatch this agent to verify that an odsp-web change actually works, after the implementer has built and started the dev server. Dispatched by the agentow skill at the Verify step (and re-dispatched each fix cycle). Returns PASS/FAIL with specific, actionable blockers. Delegate to this
plannerProactively dispatch this agent to research the odsp-web codebase for a feature or bug before writing any code. Dispatched by the agentow skill at the Research step. Returns a structured findings report: root cause (for bugs), files to change, existing patterns to follow, and a UI surface trace for
reviewerProactively dispatch this agent for an evidence-backed pre-PR code review of an odsp-web change. Dispatched after verification and before PR creation. Performs a risk inventory plus an adversarial second pass and returns a machine-validated verdict. It does NOT fix code — it reviews and reports. - v
owruns: node ${CLAUDE_PLUGIN_ROOT}/ts/dist/ow/index.js mcp
Hand a task to a cloud agent that works in its own cloud computer and returns a pull request, a report, files or an answer; start, wait for, reply to and stop tasks through a hosted MCP server with OAuth sign-in.
6 skills · 1 MCP server
Locally-hosted Codespace equivalent for codeArbiter. Pulls an untrusted repo into an ephemeral, isolated Docker container with no host-filesystem access and configurable egress, caches dependencies by content hash, then tears the box down. Requires Docker; uses installed nixpacks when available, with a generated-Dockerfile fallback.
2 skills · 5 commands
Teaches Claude to write, test and debug sandboxed rush plugins: tools, subagents, memory and agents of their own.
1 skill
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.