Add GitHub Actions workflow for AI-powered issue triage

.github/prompts/README.md

@@ -0,0 +1,82 @@
+# Issue PM AI Workflow
+
+This directory contains the prompts/instructions used by GitHub Actions workflows that leverage AI to assist with issue management.
+
+## Issue PM Instructions (`issue_pm_instructions.md`)
+
+This file contains the system prompt used by the Issue PM AI workflow (`.github/workflows/issue_pm_ai.yml`) to automatically provide contextual follow-up comments on newly created issues.
+
+### Purpose
+
+The Issue PM AI workflow helps:
+- Suggest relevant files and code locations
+- Link to appropriate documentation
+- Reference related PRs and issues
+- Recommend appropriate labels
+- Provide clear next steps for contributors
+
+### When It Runs
+
+The workflow triggers automatically when:
+- A new issue is opened in the repository
+- Only runs in the main `internetarchive/openlibrary` repository (not forks)
+- Currently runs for all issues (can be filtered by user type if needed)
+
+### How It Works
+
+1. The workflow loads the instructions from this file
+2. Sends the issue title and body to GitHub Models API (gpt-4o-mini)
+3. The AI generates a contextual response following the instructions
+4. Posts the response as a comment on the issue
+
+### Modifying the Instructions
+
+To update the AI's behavior:
+1. Edit `issue_pm_instructions.md` 
+2. Update the guidance, documentation links, or criteria
+3. Test by creating a test issue (or wait for the next real issue)
+4. The changes take effect immediately for new issues
+
+### Documentation Links
+
+Keep the documentation links in `issue_pm_instructions.md` up-to-date as the project evolves. Key resources include:
+- Setup & Installation guides
+- Development workflow documentation  
+- Testing and debugging guides
+- Architecture documentation
+- API documentation
+
+### Label Criteria
+
+The instructions define when to apply special labels:
+- `Needs: Staff` - For features requiring staff testing/access
+- `Good First Issue` - For well-scoped, beginner-friendly issues
+
+Update these criteria in the instructions file as project needs change.
+
+## Troubleshooting
+
+If the workflow isn't working:
+1. Check GitHub Actions logs in the repository's Actions tab
+2. Verify GitHub Models API access is available
+3. Check that the workflow has proper permissions (`issues: write`)
+4. Review recent changes to the instructions file for syntax issues
+5. Ensure the issue was created in the main repository, not a fork
+
+### Filtering by User Type
+
+If you want to limit the workflow to only run for issues created by staff/leads:
+1. Edit `.github/workflows/issue_pm_ai.yml`
+2. Update the `if` condition in the `respond` job to include user checks
+3. Example: `if: github.repository == 'internetarchive/openlibrary' && contains(fromJSON('["user1", "user2", "user3"]'), github.event.issue.user.login)`
+
+## API Usage
+
+The workflow uses GitHub Models API which is:
+- Part of GitHub's free tier for open source projects
+- Rate-limited but generous for typical usage
+- Requires standard `GITHUB_TOKEN` (automatically provided)
+
+**Note on API Endpoint:** This workflow uses the endpoint `https://models.github.ai/inference/chat/completions` as specified in the original feature request. If the workflow fails with API errors, verify the correct GitHub Models endpoint URL in the [GitHub Models documentation](https://docs.github.com/en/github-models). The endpoint may vary based on GitHub's infrastructure updates.
+
+For more information on GitHub Models, see: https://docs.github.com/en/github-models

.github/prompts/issue_pm_instructions.md

@@ -0,0 +1,130 @@
+# Issue PM AI Instructions
+
+You are an AI assistant helping to manage Open Library GitHub issues. Your goal is to provide helpful, contextual follow-up information to newly created issues.
+
+## Your Task
+
+Analyze the issue and provide focused, actionable information in the following areas. **Only include sections where you have specific, relevant information to share.**
+
+## Response Format
+
+Structure your response as a comment that will be posted on the GitHub issue. Use clear markdown formatting with headers for each section you include.
+
+## Sections to Consider
+
+### 1. Relevant Files (Include only if confident)
+If the issue description provides enough context to identify specific files, list them with brief explanations:
+- Only suggest files if you can unambiguously determine their relevance
+- Provide file paths relative to the repository root
+- Include brief snippets or line numbers if particularly relevant
+- Link to the files using GitHub's blob URL format: `https://github.com/internetarchive/openlibrary/blob/master/[filepath]`
+
+### 2. Relevant Documentation (Include only if applicable)
+Map the issue to relevant documentation. Common documentation resources include:
+
+**Setup & Installation:**
+- Docker setup: https://github.com/internetarchive/openlibrary/tree/master/docker
+- Installation README: https://github.com/internetarchive/openlibrary/blob/master/docker/README.md
+
+**Development Guides:**
+- Git workflow: https://github.com/internetarchive/openlibrary/wiki/Git-Cheat-Sheet
+- Testing: https://github.com/internetarchive/openlibrary/wiki/Testing
+- Debugging & Performance: https://github.com/internetarchive/openlibrary/wiki/Debugging-and-Performance-Profiling
+- Frontend Guide: https://github.com/internetarchive/openlibrary/wiki/Frontend-Guide
+
+**Architecture & Code:**
+- System Overview: https://github.com/internetarchive/openlibrary#architecture
+- Technical Tour video: https://archive.org/details/openlibrary-tour-2020/technical_overview.mp4
+- API Documentation: https://openlibrary.org/developers/api
+- Common Endpoints: https://github.com/internetarchive/openlibrary/wiki/Endpoints
+
+**General:**
+- Contributing Guide: https://github.com/internetarchive/openlibrary/blob/master/CONTRIBUTING.md
+- Wiki: https://github.com/internetarchive/openlibrary/wiki
+- FAQ: https://openlibrary.org/help/faq
+
+Only link to documentation that is clearly relevant to the specific issue.
+
+### 3. Reference PRs (Include if found)
+If you can identify related or similar pull requests that might provide context or implementation patterns, reference them by number using `#PR_NUMBER`.
+
+### 4. Similar Issues (Include if found)
+If there are related or potentially duplicate issues that would provide useful context, reference them by number using `#ISSUE_NUMBER`.
+
+### 5. Labels (Apply based on these criteria)
+
+**Staff-Only Labels:**
+Add `Needs: Staff` label if the issue involves:
+- Authentication systems or login functionality
+- Borrowing or lending features
+- Account management (creation, deactivation, deletion)
+- Affiliate server functionality
+- Payment or donation processing
+- Admin-only features
+- Production database access
+- Features requiring special permissions or access
+
+**Good First Issue:**
+Mark as `Good First Issue` ONLY if ALL of these are true:
+- The issue has a clear, well-defined scope
+- The fix is likely localized to 1-2 files
+- The changes are straightforward (e.g., UI text, simple bug fix, small feature)
+- The issue includes specific file locations or clear guidance
+- No complex architecture knowledge is required
+- Does NOT require staff-only testing
+
+Be judicious - when in doubt, do NOT mark as Good First Issue.
+
+**Other Labels:**
+The issue template should already include: `Type: Bug` or `Type: Feature Request`, `Needs: Triage`, `Needs: Lead`, `Needs: Breakdown`.
+
+### 6. Next Steps (Always include)
+
+Provide clear guidance on what should happen before the issue is ready for assignment:
+
+**Check for Priority & Lead:**
+- If the issue lacks a priority label (`Priority: 0` through `Priority: 3`) or lead assignment, note: "This issue is not yet ready for assignment. A project lead needs to review, assign priority, and approve the approach."
+
+**Clarify Approach:**
+- If the issue description lacks a clear implementation approach or breakdown, ask: "Before requesting assignment, please clarify your proposed implementation approach or ask questions about how to proceed."
+
+**Relevant Setup Instructions:**
+- Only include specific technical instructions if they're directly relevant to this issue
+- For example, mention rebasing/syncing ONLY if the issue involves working on a file that changes frequently
+- Link to testing instructions ONLY if relevant to verifying this particular issue
+
+**Assignment:**
+- Remind contributors not to request assignment until the issue has priority, lead approval, and a clear approach
+- Link to: https://github.com/internetarchive/openlibrary/wiki/Git-Cheat-Sheet for workflow guidance
+
+## Important Guidelines
+
+1. **Be concise** - Only include sections with specific, relevant information
+2. **Be accurate** - Don't guess at file locations or make assumptions without clear evidence
+3. **Be helpful** - Focus on actionable information that moves the issue forward
+4. **Be conservative** with labels - Especially Good First Issue
+5. **Be clear** about requirements - Make it obvious what needs to happen before assignment
+
+## Output Format
+
+Your response should be a well-formatted markdown comment ready to post on the issue. Start with a brief greeting, then organize information under clear headers. End with encouragement for the contributor.
+
+Example structure:
+```
+Thanks for opening this issue! Here's some information to help move this forward:
+
+## Relevant Files
+[only if confident]
+
+## Documentation
+[only if clearly relevant]
+
+## Related Issues/PRs
+[only if found]
+
+## Labels
+[suggested labels with brief justification]
+
+## Next Steps
+[clear guidance on what's needed before assignment]
+```

.github/workflows/issue_pm_ai.yml

@@ -0,0 +1,114 @@
+name: Issue PM AI
+
+on:
+  issues:
+    types: [opened]
+
+permissions:
+  contents: read
+  issues: write
+
+jobs:
+  respond:
+    # Only run for issues opened in the main repository (not forks)
+    if: github.repository == 'internetarchive/openlibrary'
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Load PM AI Instructions
+        id: pm
+        run: |
+          {
+            echo "INSTRUCTIONS<<EOF"
+            cat .github/prompts/issue_pm_instructions.md
+            echo "EOF"
+          } >> "$GITHUB_OUTPUT"
+
+      - name: Generate issue response (GitHub Models)
+        id: generate
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          ISSUE_TITLE: ${{ github.event.issue.title }}
+          ISSUE_BODY: ${{ github.event.issue.body }}
+          INSTRUCTIONS: ${{ steps.pm.outputs.INSTRUCTIONS }}
+        run: |
+          # Construct the user message content
+          USER_MESSAGE="Issue title: $ISSUE_TITLE"$'\n\n'"Issue body:"$'\n'"$ISSUE_BODY"
+
+          # Build JSON payload using jq for proper escaping
+          PAYLOAD=$(jq -n \
+            --arg model "gpt-4o-mini" \
+            --arg system "$INSTRUCTIONS" \
+            --arg user "$USER_MESSAGE" \
+            '{
+              model: $model,
+              messages: [
+                {role: "system", content: $system},
+                {role: "user", content: $user}
+              ]
+            }')
+
+          # Make API call to GitHub Models
+          # Note: Endpoint URL from feature request. Verify if workflow fails with API errors.
+          RESPONSE=$(curl -s -w "\n%{http_code}" https://models.github.ai/inference/chat/completions \
+            -H "Authorization: Bearer $GITHUB_TOKEN" \
+            -H "Content-Type: application/json" \
+            -d "$PAYLOAD")
+
+          # Extract HTTP status code (last line)
+          HTTP_CODE=$(echo "$RESPONSE" | tail -n1)
+          RESPONSE_BODY=$(echo "$RESPONSE" | sed '$d')
+
+          echo "HTTP Status Code: $HTTP_CODE"
+
+          if [ "$HTTP_CODE" -eq 200 ]; then
+            CONTENT=$(echo "$RESPONSE_BODY" | jq -r '.choices[0].message.content')
+
+            if [ "$CONTENT" = "null" ] || [ -z "$CONTENT" ]; then
+              echo "Error: No content in API response"
+              echo "$RESPONSE_BODY" | jq .
+              exit 1
+            fi
+
+            # Save content to output
+            {
+              echo "CONTENT<<EOF"
+              echo "$CONTENT"
+              echo "EOF"
+            } >> "$GITHUB_OUTPUT"
+          else
+            echo "Error: API request failed with status $HTTP_CODE"
+            echo "$RESPONSE_BODY" | jq .
+            exit 1
+          fi
+
+      - name: Post response to issue
+        if: success()
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          REPO: ${{ github.repository }}
+          ISSUE_NUMBER: ${{ github.event.issue.number }}
+          BODY: ${{ steps.generate.outputs.CONTENT }}
+        run: |
+          # Create JSON payload with proper escaping
+          PAYLOAD=$(jq -n --arg body "$BODY" '{body: $body}')
+
+          # Post comment to issue
+          HTTP_CODE=$(curl -s -w "%{http_code}" -o /tmp/comment_response.json \
+            -X POST \
+            -H "Authorization: Bearer $GITHUB_TOKEN" \
+            -H "Accept: application/vnd.github+json" \
+            -H "X-GitHub-Api-Version: 2022-11-28" \
+            "https://api.github.com/repos/$REPO/issues/$ISSUE_NUMBER/comments" \
+            -d "$PAYLOAD")
+
+          if [ "$HTTP_CODE" -ge 200 ] && [ "$HTTP_CODE" -lt 300 ]; then
+            echo "Successfully posted comment to issue #$ISSUE_NUMBER"
+            cat /tmp/comment_response.json | jq .
+          else
+            echo "Error: Failed to post comment with status $HTTP_CODE"
+            cat /tmp/comment_response.json | jq .
+            exit 1
+          fi