Add OpenClaw deployment agent prompt with checklist gates, escalation protocol, and audit requirements

[Copilot]

This PR adds a dedicated deployment-agent prompt for OpenClaw to codify a repeatable deploy process and explicit failure-handling expectations for both automation and operators. It captures end-to-end deployment stages, startup-failure remediation/escalation, and auditable execution requirements.

• What’s added

  • New prompt file: prompts/DEPLOYMENT_AGENT_PROMPT.md
  • Defines a canonical deployment checklist across:
    • pre-deploy
    • build
    • infrastructure readiness
    • secrets/config
    • deploy
    • validation
    • handoff
    • fallback/rollback

• Automation and control-flow requirements

  • Requires automation to use the checklist as source of truth.
  • Enforces CI gating before deploy.
  • Requires immediate halt on gate/step failure.
  • Requires escalation when blocked and not auto-remediable.

• OpenClaw startup-failure protocol

  • Adds explicit runbook-style flow for “fails to start/readiness”:
    • stop rollout
    • collect logs/events/exit and probe data
    • notify on-call + service owner with impact and evidence
    • perform approved remediation only
    • rollback within recovery window if not restored
    • document incident handoff

• Operator clarity and auditability

  • States checklist scope clearly: prevents process misses; does not auto-fix code/infra defects.
  • Requires diagnosable escalation for real runtime errors.
  • Requires observable/auditable records for all actions and explicit rationale + approver for deviations.

## Remediation and Escalation: OpenClaw Fails to Start

1. Stop progression immediately (no further rollout).
2. Collect evidence (logs, orchestrator events, exit/restart/probe errors).
3. Alert on-call + service owner with impact and evidence links.
4. Attempt safe remediation only if pre-approved runbook exists.
5. Roll back if startup is not restored within recovery window.

Original prompt

Add a deployment agent prompt for OpenClaw
defining:

• a reliable, repeatable checklist for deploy (pre-deploy, build, infra, secrets, deploy, validation, handoff, fallback)
• automation steps (reference checklist, CI, halt on failure, escalate if blocked)
• remediation/escalation protocol if OpenClaw fails to start, including log collection, alerting, and rollback
• operator notes clarifying that checklist prevents issues but does not auto-fix bugs; real errors must be diagnosed and escalated
• requirement for observable/auditable actions with explanations for deviations

This agent prompt should be saved as 'prompts/DEPLOYMENT_AGENT_PROMPT.md', ensuring clear and actionable steps for both automated and human operators in the repo.

The following is the prior conversation context from the user's chat exploration (may be truncated):

User: WIN for today is getting openclaw deployed today because everyones waiting on her.

Assistant: Got it—your main goal (WIN) for today is getting OpenClaw deployed, since the team is waiting on it.

Would you like a deployment checklist, the current blocking issues/open PRs for OpenClaw, help diagnosing errors, or something else specific to accelerate the deployment? Let me know what’s holding things up or how I can make this happen faster!

User: just give me checklist

Assistant: Here’s a clean deployment checklist for getting OpenClaw (or a similar system component) deployed, based on best practices and your repo’s operational playbooks:

---

OpenClaw Deployment Checklist

1. Pre-Deployment

   • Confirm latest tested version/commit is ready for deploy
   • Review and validate all environment variables and secrets required
   • Check OpenClaw system configuration (e.g., config.yaml, .env)
   • Validate code health: pass tests, lint, security checks

2. Docker & Build (if containerized)

   • Build Docker image with up-to-date code
   • Tag image appropriately (latest or version tag)
   • Push Docker image to appropriate registry (GHCR or other)

3. Infrastructure Preparation

   • Confirm destination environment is available (server, cluster, etc.)
   • Validate network/firewall and HTTPS/HTTP routes are open as needed
   • Confirm storage volumes and data directories are mounted and available

4. Secrets & Config

   • Store all necessary secrets safely (use Vault, GH Secrets, Kubernetes secrets, etc.)
   • Double-check API keys, DB strings, credentials in deployment context

5. Deploy

   • Pull latest image/tag to target host/environment
   • Deploy container(s) or start service using orchestrator (docker-compose, k8s, systemd, etc.)
   • Pass all required environment variables
   • Set up required healthchecks/probes

6. Post-Deployment Validation

   • Check service health endpoints (e.g., /health)
   • Confirm primary endpoints are accessible
   • Review logs for startup errors and warnings
   • Run a basic functionality test (submit sample intake, etc.)

7. Operational Handoff

   • Communicate deployment status to team/stakeholders
   • Document new endpoints, ports, any breaking config changes
   • Set up alerting/monitoring if needed

8. Fallback/Recovery

   • Ensure rollback plan is documented and ready (previous stable image or config)
   • Verify backups exist for any data that could change

---

If you need this checklist tailored to OpenClaw’s specific architecture or as a deploy.sh script, let me know!

User: here is the yml to review, do you feel theres a better plan or stay with your original:

OpenClaw Deployment Checklist
Pre-Deployment

Confirm latest tested version/commit is ready for deploy
Review and validate all environment variables and secrets required
Check OpenClaw system configuration (e.g., config.yaml, .env)
Validate code health: pass tests, lint, security checks
Docker & Build (if containerized)

Build Docker image with up-to-date code
Tag image appropriately (latest or version tag)
Push Docker image to appropriate registry (GHCR or other)
Infrastructure Preparation

Confirm destination environment is available (server, cluster, etc.)
Validate network/firewall and HTTPS/HTTP routes are open as needed
Confirm storage volumes and data directories are mounted and available
Secrets & Config

Store all necessary secrets safely (use Vault, GH Secrets, Kubernetes secrets, etc.)
Double-check API keys, DB strings, credentials in deployment context
Deploy

Pull latest image/tag to target host/environment
Deploy container(s) or start service using orchestrator (docker-compose, k8s, systemd, etc.)
Pass all required environment variables
Set up required healthchecks/probes
Post-Deployment Validation

Check service health endpoints (e.g., /health)
Confirm primary endpoints are accessible
Review logs for startup errors and warnings
Run a basic functionality test (submit sample intake, etc.)
Op...

This pull request was created from Copilot chat.

[erikhinla]

proceed