# Send Buildwright to your agent Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually. ## Fast path - Download the package from Yavira. - Extract it into a folder your agent can access. - Paste one of the prompts below and point your agent at the extracted folder. ## Suggested prompts ### New install ```text I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete. ``` ### Upgrade existing ```text I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run. ``` ## Machine-readable fields ```json { "schemaVersion": "1.0", "item": { "slug": "buildwright", "name": "Buildwright", "source": "tencent", "type": "skill", "category": "开发工具", "sourceUrl": "https://clawhub.ai/raunakkathuria/buildwright", "canonicalUrl": "https://clawhub.ai/raunakkathuria/buildwright", "targetPlatform": "OpenClaw" }, "install": { "downloadUrl": "/downloads/buildwright", "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=buildwright", "sourcePlatform": "tencent", "targetPlatform": "OpenClaw", "packageFormat": "ZIP package", "primaryDoc": "SKILL.md", "includedAssets": [ "SKILL.md" ], "downloadMode": "redirect", "sourceHealth": { "source": "tencent", "slug": "buildwright", "status": "healthy", "reason": "direct_download_ok", "recommendedAction": "download", "checkedAt": "2026-05-01T04:11:03.226Z", "expiresAt": "2026-05-08T04:11:03.226Z", "httpStatus": 200, "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=buildwright", "contentType": "application/zip", "probeMethod": "head", "details": { "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=buildwright", "contentDisposition": "attachment; filename=\"buildwright-0.0.9.zip\"", "redirectLocation": null, "bodySnippet": null, "slug": "buildwright" }, "scope": "item", "summary": "Item download looks usable.", "detail": "Yavira can redirect you to the upstream package for this item.", "primaryActionLabel": "Download for OpenClaw", "primaryActionHref": "/downloads/buildwright" }, "validation": { "installChecklist": [ "Use the Yavira download entry.", "Review SKILL.md after the package is downloaded.", "Confirm the extracted package contains the expected setup assets." ], "postInstallChecks": [ "Confirm the extracted package includes the expected docs or setup files.", "Validate the skill or prompts are available in your target agent workspace.", "Capture any manual follow-up steps the agent could not complete." ] } }, "links": { "detailUrl": "https://openagent3.xyz/skills/buildwright", "downloadUrl": "https://openagent3.xyz/downloads/buildwright", "agentUrl": "https://openagent3.xyz/skills/buildwright/agent", "manifestUrl": "https://openagent3.xyz/skills/buildwright/agent.json", "briefUrl": "https://openagent3.xyz/skills/buildwright/agent.md" } } ``` ## Documentation ### Buildwright Spec-driven autonomous development. Humans approve intent; agents handle everything else. ### What this skill does When activated, Buildwright directs the agent to: Read your codebase and steering documents Write a one-page spec (docs/specs/[feature]/spec.md) Stop for human approval — unless BUILDWRIGHT_AUTO_APPROVE=true Implement the feature with TDD Run quality gates: typecheck, lint, test, build Run optional security scans (if semgrep / gitleaks / trufflehog are installed) Run a Staff Engineer prompt-based code review Commit, push, and open a PR via gh ### Credentials (required) CredentialPurposeScopeHow to provideGITHUB_TOKENPush commits and open PRs via ghrepo scope (read/write)export GITHUB_TOKEN=ghp_... or configure in OpenClaw config under skills.entries.buildwright.apiKey The token must have repo scope to push branches and create pull requests. For minimal privilege, use a fine-grained personal access token scoped to a single repository with "Contents: Read and write" and "Pull requests: Read and write" permissions. Alternatively, if you use SSH for git push, the GITHUB_TOKEN is still needed for gh pr create. You can use gh auth login to authenticate the GitHub CLI separately. ### Binaries (required) BinaryPurposegitCommits and pushesghOpens PRs via GitHub CLI ### Optional tools BinaryPurposesemgrepSAST security scangitleaks / trufflehogSecrets detection ### Agent Personas (prompt-based, no binaries) Staff Engineer and Security Engineer are prompt-engineering personas — instructions loaded from .buildwright/agents/ files in the workspace. They are not external tools or binaries. The agent adopts these personas to review specs and code using defined criteria and confidence thresholds. These files contain only prompt instructions and review checklists — no secrets or credentials. ### BUILDWRIGHT_AUTO_APPROVE (optional, not a credential) This is an optional boolean flag that controls whether the agent waits for human approval at the spec stage. It is not a secret and not declared in requires.env because it is not required to run the skill. ValueBehaviorNot setInteractive (default) — stops and waits for "approved" before buildingfalseInteractive — same as defaulttrueAutonomous — commits spec to git (audit trail) and proceeds without waiting Recommendation for first use: Leave BUILDWRIGHT_AUTO_APPROVE unset until you have reviewed a few specs and are comfortable with the workflow. Start with interactive mode in a sandbox repository to observe behavior before enabling autonomous commits and PRs. ### /bw-new-feature Full pipeline for new features. Auto-detects greenfield vs existing projects. /bw-new-feature "Add OAuth2 login" Flow: Detect (greenfield or existing?) → Research → Spec → Staff Engineer validates → Human approves → TDD build → Verify → Security scan → Code review → PR Artifacts produced: docs/specs/[feature]/research.md — what the agent found in your codebase docs/specs/[feature]/spec.md — implementation plan with approaches considered ### /bw-claw Multi-agent pipeline using the Claw Architecture. Architect decomposes the feature into domain-specific claw tasks (UI, API, DB), defines interface contracts, and coordinates execution. /bw-claw "Add profile photo upload for team members" Flow: Architect analyzes → Decomposes into claw tasks → Defines interface contract → Claws execute per domain (TDD) → Architect integrates → Buildwright quality gates → PR Best for: Features that cross domain boundaries (e.g., need DB schema + API endpoint + UI component). Artifacts produced: docs/specs/[feature]/claw-plan.md — decomposition plan with interface contracts docs/specs/[feature]/claw-[domain].md — per-claw execution report ### /bw-quick Fast path for bug fixes and small tasks (<2 hrs). No spec, no approval step. /bw-quick "Fix the login timeout bug" ### /bw-ship [message] Quality pipeline for existing work: verify → security → review → PR. /bw-ship "feat(auth): add OAuth2 support" ### /bw-verify Quick checks only: typecheck → lint → test → build. ### /bw-analyse Analyse an existing codebase and write structured docs to .buildwright/codebase/. Updates tech.md with the discovered stack and commands. Run this first on any brownfield project to give every subsequent session real context. /bw-analyse Produces: STACK.md, ARCHITECTURE.md, CONVENTIONS.md, CONCERNS.md under .buildwright/codebase/. ### /bw-help Show all available commands. ### Failure Behavior If any gate fails after retries, the agent commits completed work, pushes, and opens a PR with a structured failure report. It does not leave orphaned branches or silent failures. ### Retry Policy GateRetriesRationaleVerify (typecheck, lint, test, build)2xFixable by the agentSecurity scanNoneRequires human judgmentCode reviewNoneArchitectural decisions need humans ### Security Considerations This skill performs autonomous code changes, commits, and pull requests. Understand what it does before enabling it on repositories with sensitive or production code. What the skill reads: Your repository source code, .buildwright/agents/ persona files (prompt instructions only, no secrets), and .buildwright/steering/ context files. What the skill writes: Spec files under docs/specs/, source code changes, git commits on feature branches, and pull requests via gh. What the skill does NOT do: It does not modify .env files, access secrets stores, run destructive git operations (force push, reset), or merge PRs. All changes go to feature branches with PRs for human review. Recommended setup for first use: Start with a fork or sandbox repository, not production code Leave BUILDWRIGHT_AUTO_APPROVE unset (interactive mode) to review specs before builds Use a fine-grained GitHub token scoped to a single repository with minimal permissions Rotate tokens regularly and revoke when no longer needed Review generated PRs before merging — the skill creates PRs, it does not merge them ### More Information Full documentation, source code, and setup instructions: https://github.com/raunakkathuria/buildwright ## Trust - Source: tencent - Verification: Indexed source record - Publisher: raunakkathuria - Version: 0.0.3 ## Source health - Status: healthy - Item download looks usable. - Yavira can redirect you to the upstream package for this item. - Health scope: item - Reason: direct_download_ok - Checked at: 2026-05-01T04:11:03.226Z - Expires at: 2026-05-08T04:11:03.226Z - Recommended action: Download for OpenClaw ## Links - [Detail page](https://openagent3.xyz/skills/buildwright) - [Send to Agent page](https://openagent3.xyz/skills/buildwright/agent) - [JSON manifest](https://openagent3.xyz/skills/buildwright/agent.json) - [Markdown brief](https://openagent3.xyz/skills/buildwright/agent.md) - [Download page](https://openagent3.xyz/downloads/buildwright)