Add issue-triage-pr workflow with ability to make PRs

docs/workflows/gh-agent-workflows.md

@@ -28,6 +28,7 @@ Each workflow directory contains an `example.yml` starter and a README for trigg
 | Workflow | Trigger | Description |
 | --- | --- | --- |
 | Issue Triage | New issues | Investigate and provide implementation plans |
+| Issue Triage (with PR) | New issues | Triage plus automatic draft PR for straightforward fixes |
 | Mention in Issue | `/ai` in issues | Answer questions, debug, create PRs |
 | Mention in PR | `/ai` in PRs | Review, fix code, push changes |
 | PR CI Detective | Failed PR checks | Diagnose failures and recommend fixes |

gh-agent-workflows/issue-triage-pr/README.md

@@ -0,0 +1,31 @@
+# Issue Triage (with PR)
+
+Investigate new issues and provide actionable triage analysis. For straightforward fixes, implement and open a draft PR.
+
+## Quick Install
+
+```bash
+mkdir -p .github/workflows && curl -sL \
+  https://raw.githubusercontent.com/elastic/ai-github-actions/v0/gh-agent-workflows/issue-triage-pr/example.yml \
+  -o .github/workflows/issue-triage-pr.yml
+```
+
+See [example.yml](example.yml) for the full workflow file.
+
+## Trigger
+
+| Event | Types |
+| --- | --- |
+| `issues` | `opened` |
+
+## Inputs
+
+| Input | Description | Required | Default |
+| --- | --- | --- | --- |
+| `additional-instructions` | Repo-specific instructions appended to the agent prompt | No | `""` |
+| `setup-commands` | Shell commands run before the agent starts | No | `""` |
+
+## Safe Outputs
+
+- `add-comment` — post triage analysis on the issue
+- `create-pull-request` — open a draft PR when a verified fix is implemented

gh-agent-workflows/issue-triage-pr/example.yml

@@ -0,0 +1,16 @@
+name: Issue Triage (with PR)
+on:
+  issues:
+    types: [opened]
+
+permissions:
+  contents: read
+  discussions: write
+  issues: write
+  pull-requests: write  # required by gh-aw compiler for add-comment (github/gh-aw#16673)
+
+jobs:
+  run:
+    uses: elastic/ai-github-actions/.github/workflows/gh-aw-issue-triage-pr.lock.yml@v0
+    secrets:
+      COPILOT_GITHUB_TOKEN: ${{ secrets.COPILOT_GITHUB_TOKEN }}

github/workflows/gh-aw-issue-triage-pr.md

@@ -0,0 +1,134 @@
+---
+description: "Investigate new issues and provide actionable triage analysis with optional PR creation"
+imports:
+  - gh-aw-fragments/elastic-tools.md
+  - gh-aw-fragments/runtime-setup.md
+  - gh-aw-fragments/formatting.md
+  - gh-aw-fragments/rigor.md
+  - gh-aw-fragments/mcp-pagination.md
+  - gh-aw-fragments/workflow-edit-guardrails.md
+  - gh-aw-fragments/messages-footer.md
+  - gh-aw-fragments/safe-output-add-comment.md
+  - gh-aw-fragments/safe-output-create-pr.md
+engine:
+  id: copilot
+  model: gpt-5.3-codex
+  concurrency:
+    group: "gh-aw-copilot-issue-triage-pr-${{ github.event.issue.number }}"
+on:
+  workflow_call:
+    inputs:
+      additional-instructions:
+        description: "Repo-specific instructions appended to the agent prompt"
+        type: string
+        required: false
+        default: ""
+      setup-commands:
+        description: "Shell commands to run before the agent starts (dependency install, build, etc.)"
+        type: string
+        required: false
+        default: ""
+      messages-footer:
+        description: "Footer appended to all agent comments and reviews"
+        type: string
+        required: false
+        default: ""
+    secrets:
+      COPILOT_GITHUB_TOKEN:
+        required: true
+  reaction: "eyes"
+concurrency:
+  group: issue-triage-pr-${{ github.event.issue.number }}
+  cancel-in-progress: true
+permissions:
+  contents: read
+  issues: read
+  pull-requests: read
+tools:
+  github:
+    toolsets: [repos, issues, pull_requests, search]
+  bash: true
+  web-fetch:
+network:
+  allowed:
+    - defaults
+    - github
+    - go
+    - node
+    - python
+    - ruby
+strict: false
+roles: [admin, maintainer, write]
+bots:
+  - "github-actions[bot]"
+timeout-minutes: 30
+steps:
+  - name: Repo-specific setup
+    if: ${{ inputs.setup-commands != '' }}
+    env:
+      SETUP_COMMANDS: ${{ inputs.setup-commands }}
+    run: eval "$SETUP_COMMANDS"
+---
+
+# Issue Triage Agent (with PR)
+
+Triage new issues in ${{ github.repository }} and provide actionable analysis with implementation plans. For straightforward fixes, implement and open a draft PR.
+
+## Context
+
+- **Repository**: ${{ github.repository }}
+- **Issue**: #${{ github.event.issue.number }} — ${{ github.event.issue.title }}
+
+## Constraints
+
+- **CAN**: Read files, search code, run tests and commands, comment on the issue, and open a draft PR for straightforward fixes.
+- This workflow is primarily for investigation and planning. Local file changes are for verification only unless you implement a fix.
+
+## Triage Process
+
+Follow these steps in order.
+
+### Step 1: Gather Context
+
+1. Call `generate_agents_md` to get the repository's coding guidelines and conventions. If this fails, continue without it.
+2. Read key repository files (README, CONTRIBUTING, etc.) to understand the project.
+3. Search for related issues and PRs (open and closed) that may be relevant. Call `issue_read` with method `get` on the most relevant issues to understand prior discussion, decisions, and whether this is a duplicate.
+
+### Step 2: Investigate the Codebase
+
+1. Read the issue description carefully to understand the request or problem.
+2. Explore the relevant parts of the codebase using `grep` and file reading.
+3. Run tests or commands in the workspace to verify reported bugs when possible:
+   - Run existing tests to confirm reported behavior
+   - Execute scripts to understand current behavior
+   - Run linters or static analysis if relevant
+   - Write small test files to validate findings
+   - Always explain what you're testing and why, and include command output in your response
+4. If the required change is small, clear, and verifiable, implement it directly and run relevant validation.
+
+### Step 3: Formulate Response
+
+Provide a response with the following sections. Be concise and actionable — no filler or praise.
+
+**Always lead with a tl;dr** — your first sentence should be the most important takeaway.
+
+**Sections:**
+
+1. **Recommendation** — A clear, specific recommendation for how to address the issue. If you cannot recommend a course of action, say so with a reason. "I don't know" is better than a wrong answer.
+
+2. **Findings** — Key facts from your investigation (related code, existing implementations, relevant issues/PRs). Use `<details>` tags for longer content.
+
+3. **Verification** — If you ran tests or commands, include the output. Use `<details>` tags.
+
+4. **Detailed Action Plan** — Step-by-step plan a developer could follow to implement the recommendation (or summary of implemented changes if you completed them). Reference specific files, functions, and line numbers. Use `<details>` tags.
+
+5. **Related Items** — Table of related issues, PRs, files, and web resources.
+
+Use `<details>` and `<summary>` tags for sections that would otherwise make the response too long. Short responses don't need collapsible sections. Your performance is judged by how accurate your findings are — do the investigation required to have high confidence. "I don't know" or "I'm unable to recommend a course of action" is better than a wrong answer.
+
+### Step 4: Post Response
+
+1. Call `add_comment` with your triage response.
+2. If you implemented a valid fix with verification, call `create_pull_request` to open a draft PR.
+
+${{ inputs.additional-instructions }}