MCP.Market

openclaw-release-validation skill

by openclaw·openclaw/openclaw·394k stars·MIT

Guide human release testing on an isolated OCM copy or approved in-place gateway, collect feedback, or refresh the release campaign.

C70/100content scan

Is the openclaw-release-validation skill safe?

Read the findings before you install it. We read 13 files in the folder on 2026-10-25.

  • highreferences/gateway-preparation.md:50

    Downloads a script and runs it in one step, so what runs is whatever that server sends that day. Common for installers, and still worth a look at the address.

    curl -fsSL https://github.com/openclaw/ocm/releases/latest/download/install.sh | bash

Install the openclaw-release-validation skill

A skill is a folder. Copy it into your agent's skills folder and the agent loads it when the task matches its description. 

git clone --depth 1 https://github.com/openclaw/openclaw.git /tmp/openclaw
mkdir -p ~/.claude/skills
cp -r /tmp/openclaw/.agents/skills/openclaw-release-validation ~/.claude/skills/openclaw-release-validation
available in every project

In the Claude apps, zip the folder and upload it from the Skills settings. The folder on GitHub

The instructions your agent would load

SKILL.md as published, without the frontmatter. Read it on GitHub

OpenClaw Release Validation

Help a human test an immutable latest-main build against a selected real Gateway's state. Automate preparation, triage, and reporting; the human drives interactive surfaces and judges quality. This skill collects feedback, not a release go/no-go decision.

Choose the mode

  • Campaign artifact: only when RELEASEVALIDATIONARTIFACT_PATH is present. Read artifact generation, write that artifact, and stop. GitHub is read-only; no skill-update request or interactive setup.
  • Update campaign: for an explicit refresh request, read installed skill check, then campaign lookup/dispatch. Dispatch the isolated runner for the selected explicit tag, verify its issue, and stop. Never analyze or edit the canonical issue yourself.
  • Validate release: default human-testing workflow below. It joins the campaign; it never creates or rewrites the issue body directly.

Human validation

  1. Read installed skill check. Introduce the run briefly: choose a Gateway, protect it with a disposable OCM copy or approve an in-place update, prepare the pinned main build, then test and review feedback.
  2. Read campaign lookup/dispatch. Require a current campaign and record its stable train and current beta. Gateway preparation pins the main test target separately.
  3. Read Gateway preparation. Discover without mutation, let the tester choose source and mode, then prepare the selected target. Never use the caller's active checkout.
  4. Only after readiness succeeds, offer local diagnostics for an isolated target, then read human testing to create the worksheet and guide one chosen surface at a time.
  5. On finish validation, read cleanup and report approval. It handles blocked upgrades too. Review the complete batch before any post and preserve the exact structured-report contract.

Before readiness is terminal, show only campaign/beta identity, source/mode choice, and upgrade progress or errors. Do not expose testing instructions or create a worksheet for a Gateway that never became ready.

Boundaries

An isolated OCM copy must prove all candidate-writable paths remain contained. Do not bypass a failed containment check with manual copying. Copied channel credentials require explicit authorization to stop their current owner and a restoration plan. In-place mode changes real state: present verified backup/snapshot and dry-run evidence before obtaining update approval. Rollback requires separate approval because newer migrations can make code-only reversion unsafe.

Keep one worksheet for a ready candidate; only tester-authored results count as tested. A blocked candidate uses its final upgrade-report draft instead. For setup, OCM, backup, build or cleanup failures, use the private tooling-feedback packet; do not turn these into candidate findings. With no candidate evaluation, clean up and report that fact without producing a candidate vote or posting batch.

Diagnostics are opt-in, disposable-target-only, loopback-only and content-off. Public feedback excludes raw logs, telemetry, paths, Gateway names, identifiers, credentials and setup/cleanup details. Apply the report's narrow environment allowlist and sanitization before showing the complete draft. finish validation does not authorize posts; existing explicit batch approval does. The Discord summary is a copy-ready handoff, never an automatic send.

More skills from openclaw/openclaw

  • A1passwordSet up and use 1Password CLI for sign-in, desktop integration, and reading or injecting secrets.
  • Aacp-routerRoute plain-language requests for Claude Code, Cursor, Copilot, OpenClaw ACP, OpenCode, Gemini CLI, Grok Build, Qwen, Kiro, Kimi, iFlow, Factory Droid, Kilocode, or explicit ACP harness work into either OpenClaw ACP runtime sessions or direct acpx-driven sessions ("telephone game" flow). For coding-agent thread requests, read this skill first, then use only `sessions_spawn` for thread creation. Codex chat binding defaults to the native Codex app-server plugin unless ACP is explicit or background spawn needs ACP.
  • Aadd-model-providerAdd and live-prove a model provider with non-interactive config one-liners, without exposing credentials.
  • Aagent-transcriptRequested GitHub PR/issue agent transcripts: redact, trim, preview, and insert safely.
  • Aapple-notesCreate, view, edit, delete, search, move, or export Apple Notes via the memo CLI on macOS.
  • Aapple-remindersList, add, edit, complete, or delete Apple Reminders and reminder lists via remindctl.
  • Aauto-qaRun OpenClaw-wide autonomous QA and live/stress campaigns across independent subsystem lanes, with verified fixes and a resumable evidence report.
  • AautoreviewStructured code review through the shared OpenClaw agent-skills installation.
  • Abear-notesCreate, search, and manage Bear notes via grizzly CLI.
  • Ablock-kitUse proactively for structured or interactive Slack replies, and when asked to author or validate native Slack Block Kit JSON.
  • AblogwatcherMonitor blogs and RSS/Atom feeds for updates using the blogwatcher CLI.
  • AblucliBluOS CLI (blu) for discovery, playback, grouping, and volume.

All agent skills → · MCP servers