Add deployment_status trigger guidance to create-agentic-workflow prompt

[Copilot]

The agent was defaulting to workflow_run + if:failure for all deployment monitoring scenarios, missing deployment_status as the correct trigger for external deployment services (Heroku, Vercel, Railway, Fly.io) that post status back to GitHub.

Changes

• .github/aw/create-agentic-workflow.md — Extended the trigger inference list with:

  • deployment_status entry for external deployment monitoring, with the correct if: condition and a reference to the new example
  • workflow_run entry for GitHub Actions-internal pipeline monitoring
  • An explicit decision rule distinguishing the two: use deployment_status for services integrating with the GitHub Deployments API; use workflow_run for in-repo Actions pipelines

• .github/aw/deployment-status.md (new) — Reference prompt (not a runnable workflow) that the agent consults when creating a deployment monitoring workflow. Formatted like report.md with only description in the frontmatter. The body includes:

  • Recommended frontmatter snippet with deployment_status trigger, read-only permissions, and safe-outputs
  • Table of available github.event.deployment_status.* and github.event.deployment.* context expressions
  • Agent instructions pattern for verifying the failure, checking duplicates, and filing an incident issue
  • Decision guide for choosing deployment_status vs workflow_run

Original prompt

---

This section details on the original issue you should resolve

<issue_title>[plan] Add deployment_status trigger guidance to create-agentic-workflow prompt</issue_title>
<issue_description>## Objective

Add deployment_status event documentation and examples to the create-agentic-workflow prompt so the agent suggests it for external deployment monitoring scenarios instead of defaulting to workflow_run.

Context

From Agent Persona Exploration discussion github/gh-aw#18250, the DO-1 (Failed Deployment Incident Creator) scenario scored 4.4/5.0. The agent applies the workflow_run + if:failure pattern correctly but misses deployment_status as a better trigger for external deployment services (e.g., Heroku, Vercel, Railway, Fly.io).

Approach

1. Locate the create-agentic-workflow prompt/skill
2. Add a section on deployment_status event:
   • Explain when to use it: external deployment services that post deployment status back to GitHub
   • Show the trigger syntax:

     on:
       deployment_status:

   • Add if: github.event.deployment_status.state == 'failure' condition example
3. Distinguish from workflow_run: use deployment_status for external deployments, workflow_run for GitHub Actions-internal pipelines
4. Consider adding a reference workflow for deployment incident creation

Files to Modify

• Update: the create-agentic-workflow prompt/skill with deployment_status guidance
• Optionally create: .github/workflows/deployment-incident.md reference workflow

Acceptance Criteria

• deployment_status trigger is documented in the create prompt
• Agent suggests deployment_status when user describes external deployment monitoring
• Clear guidance on when to choose deployment_status vs workflow_run

Generated by Plan Command for issue #discussion #18250

• expires on Feb 27, 2026, 4:46 AM UTC

Comments on the Issue (you are @copilot in this section)

@pelikhan Use .github/aw/deployment-status.md

• Fixes [plan] Add deployment_status trigger guidance to create-agentic-workflow prompt #18261

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[pelikhan]

@copilot no write permissions

[Copilot]

@copilot no write permissions

Fixed in f3ff02a — changed issues: write to issues: read. The safe-outputs system handles the actual write operation, so the workflow only needs read permissions.

[pelikhan]

@copilot format this file as a prompt, not an agentic workflow

[Copilot]

Done in 4913808 — reformatted as a prompt with just description in the frontmatter. The file now follows the same pattern as report.md and github-agentic-workflows.md: reference/instructional content the agent consults when creating a deployment_status workflow, including the recommended frontmatter snippet, available event expressions, an agent instructions template, and the deployment_status vs workflow_run decision guide.

[Copilot]

Pull request overview

Updates the agentic workflow creation guidance so external deployment failure monitoring uses the deployment_status trigger (instead of defaulting to workflow_run) and adds a reference prompt for deployment-status-based incident workflows.

Changes:

• Extend trigger inference guidance in create-agentic-workflow.md to distinguish deployment_status (external deploy services) vs workflow_run (in-repo pipelines).
• Add .github/aw/deployment-status.md as a reference pattern for deployment failure monitoring and incident issue creation.
• Update the dispatcher agent doc to include routing for the test-coverage prompt.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.

File	Description
.github/aw/deployment-status.md	Adds a reference pattern (frontmatter + context expressions + instruction template) for deployment_status-driven incident creation.
.github/aw/create-agentic-workflow.md	Adds explicit trigger-selection guidance for external deployments vs internal workflow monitoring.
.github/agents/agentic-workflows.agent.md	Documents a new dispatcher route/use-case section for test coverage analysis workflows.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The markdown table is using || at the start of rows, which won’t render as a table in standard Markdown. Use single | delimiters (like other docs in this repo) so the expressions/context list renders correctly.

[Copilot]

This guidance filters only on github.event.deployment_status.state == 'failure', but deployment statuses may also report unsuccessful deployments as error. To avoid missing failures from external providers, consider updating the recommended condition to include both failure and error (or a non-success check).

Suggested change

	- **External deployment monitoring** (Heroku, Vercel, Railway, Fly.io, etc.) → `on: deployment_status:` with `if: ${{ github.event.deployment_status.state == 'failure' }}` — use this when third-party services post deployment status back to GitHub. See reference: @.github/aw/deployment-status.md
	- **External deployment monitoring** (Heroku, Vercel, Railway, Fly.io, etc.) → `on: deployment_status:` with `if: ${{ github.event.deployment_status.state == 'failure' || github.event.deployment_status.state == 'error' }}` — use this when third-party services post deployment status back to GitHub. See reference: @.github/aw/deployment-status.md

[Copilot]

The recommended if: filter is indented with a leading space, which makes the YAML snippet invalid if copied verbatim. Also, deployment_status.state can be error as well as failure for unsuccessful deployments, so filtering only on 'failure' can miss real deploy failures—consider matching both (failure or error) or using a broader non-success check.