diff --git a/.deepwork/jobs/pr_docs_update/AGENTS.md b/.deepwork/jobs/pr_docs_update/AGENTS.md
new file mode 100644
index 00000000..3650e61b
--- /dev/null
+++ b/.deepwork/jobs/pr_docs_update/AGENTS.md
@@ -0,0 +1,70 @@
+# Project Context for pr_docs_update
+
+This job automatically updates documentation by analyzing recently merged pull requests.
+
+## Purpose
+
+The `pr_docs_update` job helps maintain up-to-date AGENTS.md files and copilot-specific instructions by:
+- Fetching recently merged PRs from the GitHub repository
+- Analyzing PR content for documentation-relevant changes
+- Updating appropriate documentation files
+- Tracking processed PRs to enable incremental updates
+
+## Dual Location Maintenance
+
+**Important**: This is an example job that can be used in any repository. If you're working on the DeepWork repository itself and want to make this a standard job, follow the dual location pattern:
+
+1. **Source of truth**: `src/deepwork/standard_jobs/pr_docs_update/`
+   - Make changes here first
+   - Tracked in version control
+
+2. **Working copy**: `.deepwork/jobs/pr_docs_update/`
+   - Updated from source using `deepwork install`
+   - Used by `deepwork sync` to generate commands
+
+## File Organization
+
+```
+pr_docs_update/
+├── AGENTS.md              # This file
+├── job.yml                # Job definition
+└── steps/
+    ├── fetch_prs.md       # Fetch merged PRs from GitHub
+    ├── analyze_prs.md     # Analyze PR content
+    └── update_docs.md     # Update documentation files
+```
+
+## How It Works
+
+1. **State Tracking**: Uses `.deepwork/pr_docs_state.json` to track processed PRs
+2. **Incremental Processing**: Only processes new PRs, skips already-processed ones
+3. **Flexible Configuration**: Can specify PR count and state file location
+4. **Multiple Runs**: Can be run repeatedly to go deeper into PR history
+
+## Usage Notes
+
+- Run periodically (e.g., weekly) to keep documentation current
+- Increase `pr_count` parameter to process more PRs at once
+- State file prevents duplicate processing across runs
+- Can be used in any repository with GitHub PRs
+
+## State File Format
+
+```json
+{
+  "processed_prs": [123, 456, 789],
+  "last_update": "2026-01-19T07:00:00Z",
+  "total_processed": 3
+}
+```
+
+## Requirements
+
+- GitHub API access (uses GitHub MCP tools)
+- Write access to documentation files
+- Git repository with GitHub remote
+
+## Last Updated
+
+- Date: 2026-01-19
+- From: Initial job creation
diff --git a/.deepwork/jobs/pr_docs_update/USAGE.md b/.deepwork/jobs/pr_docs_update/USAGE.md
new file mode 100644
index 00000000..cb6ac3c6
--- /dev/null
+++ b/.deepwork/jobs/pr_docs_update/USAGE.md
@@ -0,0 +1,162 @@
+# PR Documentation Update Job - Example Usage
+
+This document provides example usage scenarios for the `pr_docs_update` job.
+
+## Overview
+
+The `pr_docs_update` job automatically updates documentation by analyzing recently merged pull requests. It's designed to:
+- Keep AGENTS.md files up to date with code changes
+- Update copilot-specific instructions when relevant
+- Track processed PRs to avoid duplicates
+- Enable incremental processing across multiple runs
+
+## Basic Usage
+
+### First Run - Process Last 10 PRs
+
+```bash
+# Run all steps to process the 10 most recent merged PRs
+/pr_docs_update.fetch_prs
+/pr_docs_update.analyze_prs
+/pr_docs_update.update_docs
+```
+
+The job will:
+1. Fetch the 10 most recent merged PRs
+2. Filter out any already processed (from state file)
+3. Analyze each PR for documentation relevance
+4. Update appropriate documentation files
+5. Save processed PR numbers to state file
+
+### Subsequent Runs - Process More PRs
+
+```bash
+# Run with a higher count to go deeper into history
+# The state file automatically skips already-processed PRs
+/pr_docs_update.fetch_prs pr_count=20
+/pr_docs_update.analyze_prs
+/pr_docs_update.update_docs
+```
+
+## Advanced Usage
+
+### Custom State File Location
+
+If you want to track state separately for different purposes:
+
+```bash
+/pr_docs_update.fetch_prs state_file=".deepwork/pr_docs_copilot_state.json"
+/pr_docs_update.analyze_prs
+/pr_docs_update.update_docs state_file=".deepwork/pr_docs_copilot_state.json"
+```
+
+### Reprocessing Specific PRs
+
+To reprocess specific PRs, edit the state file:
+
+```bash
+# View current state
+cat .deepwork/pr_docs_state.json
+
+# Remove specific PR numbers you want to reprocess
+# Edit the file and remove those numbers from the "processed_prs" array
+```
+
+### Processing Large Batches
+
+For initial setup or catching up after a long time:
+
+```bash
+# Process 50 PRs at once
+/pr_docs_update.fetch_prs pr_count=50
+/pr_docs_update.analyze_prs
+/pr_docs_update.update_docs
+```
+
+## State File Structure
+
+The state file (`.deepwork/pr_docs_state.json`) tracks processed PRs:
+
+```json
+{
+  "processed_prs": [123, 456, 789],
+  "last_update": "2026-01-19T07:00:00Z",
+  "total_processed": 3
+}
+```
+
+## What Gets Updated?
+
+### AGENTS.md Files
+
+For general code changes that all agents should know about:
+- New jobs or workflows
+- Codebase structure changes
+- New conventions or patterns
+- Important context for working with the code
+
+Located at: `.deepwork/jobs/[job_name]/AGENTS.md`
+
+### Copilot-Specific Instructions
+
+For changes specific to GitHub Copilot integration:
+- Copilot-specific features
+- Copilot workflow changes
+- Copilot tool usage patterns
+
+Located at: `.github/copilot/` or similar copilot instruction locations
+
+## Scheduling Recommendations
+
+- **Weekly**: Run with default settings (10 PRs) to stay current
+- **After Major Releases**: Run with higher count (30-50 PRs)
+- **Initial Setup**: Process all recent PRs in batches
+
+## Integration with Existing Workflows
+
+This job can be combined with other DeepWork jobs:
+
+```bash
+# Update docs from PRs, then commit the changes
+/pr_docs_update.fetch_prs
+/pr_docs_update.analyze_prs
+/pr_docs_update.update_docs
+
+# Review and commit
+/commit.test
+/commit.lint
+/commit.commit_and_push
+```
+
+## Troubleshooting
+
+### No New PRs Found
+
+If all fetched PRs have been processed:
+- Increase the `pr_count` parameter
+- Check the state file to see what's been processed
+- Consider clearing the state file to reprocess all PRs
+
+### GitHub API Rate Limiting
+
+If you hit rate limits:
+- Reduce the `pr_count` parameter
+- Wait for the rate limit to reset
+- Use authenticated GitHub API access if available
+
+### Documentation Files Not Found
+
+If expected documentation files don't exist:
+- The job will notify you about missing files
+- Create the necessary AGENTS.md or instruction files
+- Re-run the update step
+
+## Example Output
+
+After running the job, you'll have:
+
+1. **pr_list.json**: List of fetched PRs (filtered by state)
+2. **pr_analysis.md**: Analysis of each PR with recommendations
+3. **updated_files_list.md**: Summary of what was updated
+4. **Updated documentation files**: AGENTS.md or copilot instructions with new content
+5. **Updated state file**: Tracking newly processed PRs
diff --git a/.deepwork/jobs/pr_docs_update/job.yml b/.deepwork/jobs/pr_docs_update/job.yml
new file mode 100644
index 00000000..d057b4da
--- /dev/null
+++ b/.deepwork/jobs/pr_docs_update/job.yml
@@ -0,0 +1,68 @@
+name: pr_docs_update
+version: "1.0.0"
+summary: "Update documentation from recently merged PRs"
+description: |
+  A workflow that analyzes recently merged pull requests and updates either AGENTS.md
+  or copilot-specific instructions based on the PR content. Tracks processed PRs to
+  avoid duplicates and can be run multiple times to go further back in PR history.
+
+  This job helps maintain up-to-date documentation by:
+  1. Fetching recently merged PRs from GitHub
+  2. Analyzing PR content to identify documentation-relevant changes
+  3. Updating AGENTS.md (for general context) or copilot instructions (for copilot-specific changes)
+  4. Tracking which PRs have been processed to enable incremental updates
+
+  Use this job periodically to keep documentation in sync with code changes.
+
+changelog:
+  - version: "1.0.0"
+    changes: "Initial job creation"
+
+steps:
+  - id: fetch_prs
+    name: "Fetch Recently Merged PRs"
+    description: "Retrieve list of recently merged pull requests from GitHub"
+    instructions_file: steps/fetch_prs.md
+    inputs:
+      - name: pr_count
+        description: "Number of recent PRs to fetch (default: 10)"
+      - name: state_file
+        description: "Path to state file tracking processed PRs (default: .deepwork/pr_docs_state.json)"
+    outputs:
+      - pr_list.json
+    dependencies: []
+
+  - id: analyze_prs
+    name: "Analyze PR Content"
+    description: "Analyze PR content to identify documentation-relevant changes"
+    instructions_file: steps/analyze_prs.md
+    inputs:
+      - file: pr_list.json
+        from_step: fetch_prs
+    outputs:
+      - pr_analysis.md
+    dependencies:
+      - fetch_prs
+
+  - id: update_docs
+    name: "Update Documentation"
+    description: "Update AGENTS.md or copilot instructions based on PR analysis"
+    instructions_file: steps/update_docs.md
+    inputs:
+      - file: pr_analysis.md
+        from_step: analyze_prs
+      - name: state_file
+        description: "Path to state file tracking processed PRs (default: .deepwork/pr_docs_state.json)"
+    outputs:
+      - updated_files_list.md
+    dependencies:
+      - analyze_prs
+    hooks:
+      after_agent:
+        - prompt: |
+            Verify the documentation update is complete:
+            1. Relevant documentation files (AGENTS.md or copilot instructions) were identified
+            2. Documentation was updated with insights from analyzed PRs
+            3. State file was updated with processed PR numbers
+            4. Changes are ready to be committed
+            If ALL criteria are met, include `<promise>✓ Quality Criteria Met</promise>`.
diff --git a/.deepwork/jobs/pr_docs_update/steps/analyze_prs.md b/.deepwork/jobs/pr_docs_update/steps/analyze_prs.md
new file mode 100644
index 00000000..84ac31d0
--- /dev/null
+++ b/.deepwork/jobs/pr_docs_update/steps/analyze_prs.md
@@ -0,0 +1,137 @@
+# Analyze PR Content
+
+## Objective
+
+Analyze the content of unprocessed pull requests to identify which ones contain changes relevant to documentation updates, specifically for AGENTS.md files or copilot-specific instructions.
+
+## Task
+
+Review each PR from the input list and determine what documentation updates are needed based on the PR's changes, title, description, and files modified.
+
+### Step 1: Load PR List
+
+1. **Read the input file** `pr_list.json`
+   - Parse the JSON structure
+   - Extract the list of PRs to analyze
+
+2. **Verify the data**
+   - Ensure PRs have all required fields
+   - Note the repository context
+
+### Step 2: Analyze Each PR
+
+For each PR in the list, evaluate its relevance to documentation:
+
+1. **Examine Changed Files**
+   - Look for patterns that suggest documentation impact:
+     - New features or capabilities added
+     - Changes to job definitions or workflows
+     - API or interface modifications
+     - New standard jobs or tools
+     - Configuration changes
+     - Agent or LLM interaction changes
+   
+2. **Review PR Title and Description**
+   - Identify the purpose of the PR
+   - Note any documentation mentions
+   - Look for keywords like: "add", "new", "feature", "job", "workflow", "agent", "copilot"
+
+3. **Categorize the PR**
+   - **General documentation**: Changes that affect all agents (→ AGENTS.md)
+   - **Copilot-specific**: Changes specific to GitHub Copilot integration (→ copilot instructions)
+   - **Not relevant**: Changes with no documentation impact
+   - **Unclear**: Needs further investigation
+
+4. **Extract Documentation Insights**
+   For relevant PRs, identify:
+   - What was added or changed?
+   - What should agents know about this change?
+   - Which documentation file should be updated?
+   - What specific information should be added?
+
+### Step 3: Determine Documentation Updates
+
+For each relevant PR, create a documentation recommendation:
+
+1. **For AGENTS.md updates**:
+   - New job or workflow information
+   - Codebase structure changes
+   - New conventions or patterns
+   - Important context for all agents
+
+2. **For copilot-specific instructions**:
+   - Copilot-specific features or integrations
+   - Copilot workflow changes
+   - Copilot tool usage patterns
+
+3. **Documentation entry format**:
+   - Clear, concise description
+   - Reference to the PR number
+   - Suggested location in the documentation
+   - Specific text to add or update
+
+### Step 4: Create Analysis Report
+
+Create `pr_analysis.md` with the following structure:
+
+```markdown
+# PR Documentation Analysis
+
+**Repository**: owner/repo-name
+**Analysis Date**: 2026-01-19
+**PRs Analyzed**: 5
+**Documentation Updates Needed**: 3
+
+## PRs Requiring Documentation Updates
+
+### PR #123: Add new feature
+- **Merge Date**: 2026-01-15
+- **Author**: username
+- **URL**: https://github.com/owner/repo/pull/123
+- **Category**: General Documentation (AGENTS.md)
+- **Summary**: This PR adds a new job type for automated testing
+- **Documentation Update**: 
+  - File: `.deepwork/jobs/[job_name]/AGENTS.md`
+  - Section: Job Structure
+  - Content: Document the new automated testing job available in the repository
+  - Rationale: Agents should be aware of this capability for future task planning
+
+### PR #456: Update copilot integration
+- **Merge Date**: 2026-01-14
+- **Author**: username2
+- **URL**: https://github.com/owner/repo/pull/456
+- **Category**: Copilot-Specific Instructions
+- **Summary**: Enhanced copilot workflow with new commands
+- **Documentation Update**: 
+  - File: Copilot-specific instruction files
+  - Section: Available Commands
+  - Content: Document the new `/copilot-command` syntax and usage
+  - Rationale: Copilot agents need to know about new command capabilities
+
+## PRs Not Requiring Updates
+
+### PR #789: Fix typo in README
+- **Reason**: Minor documentation fix, no agent-relevant changes
+```
+
+## Output Format
+
+- **File**: `pr_analysis.md`
+- **Content**: Markdown document with analysis results and recommendations
+- **Location**: Current working directory
+
+## Quality Criteria
+
+- ✓ All PRs from input list analyzed
+- ✓ Each PR categorized (general docs, copilot-specific, not relevant, or unclear)
+- ✓ Documentation updates clearly specified with target file and content
+- ✓ Rationale provided for each recommended update
+- ✓ Summary statistics included (PRs analyzed, updates needed)
+- ✓ Clear, actionable recommendations for the next step
+
+## Notes
+
+- Focus on changes that provide value to AI agents working with the codebase
+- Avoid documenting trivial changes (typos, formatting, etc.)
+- When in doubt, err on the side of including information that might be helpful
+- Consider the perspective of an agent encountering this codebase for the first time
diff --git a/.deepwork/jobs/pr_docs_update/steps/fetch_prs.md b/.deepwork/jobs/pr_docs_update/steps/fetch_prs.md
new file mode 100644
index 00000000..1c53ee8e
--- /dev/null
+++ b/.deepwork/jobs/pr_docs_update/steps/fetch_prs.md
@@ -0,0 +1,108 @@
+# Fetch Recently Merged PRs
+
+## Objective
+
+Retrieve a list of recently merged pull requests from GitHub and filter out any that have already been processed, based on a state file.
+
+## Task
+
+Fetch recently merged PRs from the current repository and prepare a list of unprocessed PRs for analysis.
+
+### Step 1: Understand the Parameters
+
+1. **PR Count**: Determine how many recent PRs to fetch
+   - Default: 10 PRs
+   - If user specified a count, use that value
+   - Can fetch more to go deeper into history
+
+2. **State File**: Load the state file to track processed PRs
+   - Default location: `.deepwork/pr_docs_state.json`
+   - If file doesn't exist, create it with an empty list
+   - File format: `{"processed_prs": [123, 456, 789]}`
+
+### Step 2: Fetch Merged PRs from GitHub
+
+Use GitHub tools available to you to fetch recently merged pull requests:
+
+1. **Get the repository information**
+   - Repository owner and name from current git repository
+   - Use `git remote -v` to identify the GitHub repository
+
+2. **Fetch merged PRs**
+   - Use GitHub API or available GitHub tools to list merged pull requests
+   - Sort by merge date (most recent first)
+   - Fetch the number requested (default 10, or user-specified count)
+   
+3. **For each PR, collect**:
+   - PR number
+   - Title
+   - Description/body
+   - Merge date
+   - Author
+   - Labels (if any)
+   - Changed files (summary)
+   - Link to the PR
+
+### Step 3: Filter Out Already Processed PRs
+
+1. **Load the state file**
+   - Read `.deepwork/pr_docs_state.json` (or user-specified path)
+   - Parse the JSON to get the list of processed PR numbers
+   - If file doesn't exist, treat it as empty (no processed PRs)
+
+2. **Filter the PR list**
+   - Remove PRs whose numbers are in the `processed_prs` list
+   - Keep only unprocessed PRs
+
+3. **Handle the case of no new PRs**
+   - If all fetched PRs have been processed, inform the user
+   - Suggest fetching more PRs by increasing the count
+   - Recommend using a larger `pr_count` value to go deeper into history
+
+### Step 4: Create Output File
+
+Create `pr_list.json` with the following structure:
+
+```json
+{
+  "repository": "owner/repo-name",
+  "fetch_date": "2026-01-19T07:00:00Z",
+  "total_fetched": 10,
+  "new_prs": 5,
+  "prs": [
+    {
+      "number": 123,
+      "title": "Add new feature",
+      "description": "This PR adds...",
+      "merge_date": "2026-01-15T10:30:00Z",
+      "author": "username",
+      "labels": ["enhancement", "documentation"],
+      "changed_files": ["src/feature.py", "docs/README.md"],
+      "url": "https://github.com/owner/repo/pull/123"
+    }
+  ]
+}
+```
+
+## Output Format
+
+- **File**: `pr_list.json`
+- **Content**: JSON structure with repository info and list of unprocessed merged PRs
+- **Location**: Current working directory
+
+## Quality Criteria
+
+- ✓ Repository information correctly identified from git remote
+- ✓ At least the requested number of PRs fetched (unless fewer exist)
+- ✓ State file loaded (or created if missing)
+- ✓ Already-processed PRs filtered out
+- ✓ PR data includes all required fields (number, title, description, merge_date, author, url)
+- ✓ Output file created with valid JSON structure
+- ✓ If no new PRs found, user informed about options to fetch more
+
+## Notes
+
+- This step uses GitHub API access, which may require authentication
+- If you encounter rate limiting, inform the user
+- The state file prevents duplicate processing across multiple runs
+- Users can delete or edit the state file to reprocess specific PRs if needed
diff --git a/.deepwork/jobs/pr_docs_update/steps/update_docs.md b/.deepwork/jobs/pr_docs_update/steps/update_docs.md
new file mode 100644
index 00000000..f3ba3a1e
--- /dev/null
+++ b/.deepwork/jobs/pr_docs_update/steps/update_docs.md
@@ -0,0 +1,168 @@
+# Update Documentation
+
+## Objective
+
+Update AGENTS.md files or copilot-specific instructions based on the PR analysis, and track processed PRs in the state file to prevent duplicate processing in future runs.
+
+## Task
+
+Apply the documentation recommendations from the analysis step, update the appropriate files, and maintain the state file for tracking.
+
+### Step 1: Load the Analysis
+
+1. **Read `pr_analysis.md`**
+   - Parse the recommendations
+   - Identify which files need updates
+   - Group updates by target file
+
+2. **Load the state file**
+   - Read `.deepwork/pr_docs_state.json` (or user-specified path)
+   - Prepare to add newly processed PR numbers
+
+### Step 2: Identify Target Documentation Files
+
+Based on the analysis, determine which documentation files to update:
+
+1. **AGENTS.md files**
+   - Check for existing AGENTS.md files in relevant job directories
+   - Typical locations:
+     - `.deepwork/jobs/[job_name]/AGENTS.md`
+     - `src/deepwork/standard_jobs/[job_name]/AGENTS.md`
+   - If no AGENTS.md exists for relevant jobs, consider creating one
+
+2. **Copilot-specific instructions**
+   - Look for copilot instruction files
+   - Common locations might include:
+     - `.github/copilot/`
+     - `.copilot/`
+     - Other copilot-specific configuration locations
+   - If no copilot instructions exist and updates are needed, note this for the user
+
+3. **Create an update plan**
+   - List all files to be updated
+   - Organize updates by file to minimize context switching
+
+### Step 3: Update Documentation Files
+
+For each documentation file identified:
+
+1. **Review existing content**
+   - Read the current file content
+   - Understand the existing structure and sections
+   - Identify where new content should be added
+
+2. **Apply updates**
+   - Add new information based on PR analysis
+   - Include PR reference numbers for traceability
+   - Follow the existing format and style
+   - Maintain chronological order if applicable
+   - Keep entries clear and concise
+
+3. **AGENTS.md format example**:
+   ```markdown
+   ## Last Updated
+   
+   - Date: 2026-01-19
+   - From: PR documentation sync (PRs #123, #456)
+   
+   ## Recent Changes
+   
+   ### PR #123: New Automated Testing Job
+   - A new job type for automated testing is now available
+   - Location: `.deepwork/jobs/automated_test/`
+   - Use this for test automation workflows
+   - See job.yml for full specification
+   ```
+
+4. **Handle multiple file updates**
+   - Update all relevant files identified in the analysis
+   - Maintain consistency across updates
+   - Ensure cross-references are accurate
+
+### Step 4: Update State File
+
+1. **Collect processed PR numbers**
+   - Extract PR numbers from the analysis
+   - Include all PRs (both relevant and not relevant)
+   - This ensures we don't reprocess any PRs
+
+2. **Update the state file**
+   - Load existing `processed_prs` list
+   - Add new PR numbers to the list
+   - Remove duplicates if any
+   - Sort the list (optional, for readability)
+
+3. **Write the updated state file**
+   ```json
+   {
+     "processed_prs": [123, 456, 789, 1011, 1213],
+     "last_update": "2026-01-19T07:00:00Z",
+     "total_processed": 5
+   }
+   ```
+
+### Step 5: Create Summary Report
+
+Create `updated_files_list.md` documenting what was done:
+
+```markdown
+# Documentation Update Summary
+
+**Date**: 2026-01-19
+**PRs Processed**: 5 (PRs #123, #456, #789, #1011, #1213)
+**Files Updated**: 2
+
+## Updated Files
+
+### .deepwork/jobs/deepwork_jobs/AGENTS.md
+- Added information about new automated testing job (PR #123)
+- Updated last modified date
+
+### .github/copilot/instructions.md
+- Documented new copilot command syntax (PR #456)
+- Added usage examples
+
+## State File Updated
+
+- State file: `.deepwork/pr_docs_state.json`
+- New processed PRs added: 5
+- Total PRs tracked: 15
+
+## PRs Processed But Not Documented
+
+- PR #789: Minor typo fix (not relevant)
+- PR #1011: Internal refactoring (no documentation impact)
+- PR #1213: CI configuration (no agent-visible changes)
+
+## Next Steps
+
+To process more PRs from further back in history, run this job again with:
+- Increase `pr_count` parameter (e.g., 20 or 30)
+- The state file will automatically skip already processed PRs
+```
+
+## Output Format
+
+- **File**: `updated_files_list.md`
+- **Content**: Summary of changes made and files updated
+- **Location**: Current working directory
+
+## Quality Criteria
+
+- ✓ All relevant documentation files identified and updated
+- ✓ Updates include PR references for traceability
+- ✓ Existing file structure and format maintained
+- ✓ State file updated with all processed PR numbers
+- ✓ State file includes metadata (last_update, total_processed)
+- ✓ Summary report created with clear list of changes
+- ✓ Changes are ready to be committed
+- ✓ If no files exist for updates, user is informed
+
+## Notes
+
+- If AGENTS.md files don't exist in expected locations, ask the user if they should be created
+- Maintain the existing tone and style of documentation files
+- Include enough context so future readers understand the changes
+- The state file enables incremental processing across multiple runs
+- Users can manually edit the state file to reprocess specific PRs if needed
+- Consider running a git diff to show the user what changed before committing
diff --git a/src/deepwork/standard_jobs/pr_docs_update/AGENTS.md b/src/deepwork/standard_jobs/pr_docs_update/AGENTS.md
new file mode 100644
index 00000000..3650e61b
--- /dev/null
+++ b/src/deepwork/standard_jobs/pr_docs_update/AGENTS.md
@@ -0,0 +1,70 @@
+# Project Context for pr_docs_update
+
+This job automatically updates documentation by analyzing recently merged pull requests.
+
+## Purpose
+
+The `pr_docs_update` job helps maintain up-to-date AGENTS.md files and copilot-specific instructions by:
+- Fetching recently merged PRs from the GitHub repository
+- Analyzing PR content for documentation-relevant changes
+- Updating appropriate documentation files
+- Tracking processed PRs to enable incremental updates
+
+## Dual Location Maintenance
+
+**Important**: This is an example job that can be used in any repository. If you're working on the DeepWork repository itself and want to make this a standard job, follow the dual location pattern:
+
+1. **Source of truth**: `src/deepwork/standard_jobs/pr_docs_update/`
+   - Make changes here first
+   - Tracked in version control
+
+2. **Working copy**: `.deepwork/jobs/pr_docs_update/`
+   - Updated from source using `deepwork install`
+   - Used by `deepwork sync` to generate commands
+
+## File Organization
+
+```
+pr_docs_update/
+├── AGENTS.md              # This file
+├── job.yml                # Job definition
+└── steps/
+    ├── fetch_prs.md       # Fetch merged PRs from GitHub
+    ├── analyze_prs.md     # Analyze PR content
+    └── update_docs.md     # Update documentation files
+```
+
+## How It Works
+
+1. **State Tracking**: Uses `.deepwork/pr_docs_state.json` to track processed PRs
+2. **Incremental Processing**: Only processes new PRs, skips already-processed ones
+3. **Flexible Configuration**: Can specify PR count and state file location
+4. **Multiple Runs**: Can be run repeatedly to go deeper into PR history
+
+## Usage Notes
+
+- Run periodically (e.g., weekly) to keep documentation current
+- Increase `pr_count` parameter to process more PRs at once
+- State file prevents duplicate processing across runs
+- Can be used in any repository with GitHub PRs
+
+## State File Format
+
+```json
+{
+  "processed_prs": [123, 456, 789],
+  "last_update": "2026-01-19T07:00:00Z",
+  "total_processed": 3
+}
+```
+
+## Requirements
+
+- GitHub API access (uses GitHub MCP tools)
+- Write access to documentation files
+- Git repository with GitHub remote
+
+## Last Updated
+
+- Date: 2026-01-19
+- From: Initial job creation
diff --git a/src/deepwork/standard_jobs/pr_docs_update/USAGE.md b/src/deepwork/standard_jobs/pr_docs_update/USAGE.md
new file mode 100644
index 00000000..cb6ac3c6
--- /dev/null
+++ b/src/deepwork/standard_jobs/pr_docs_update/USAGE.md
@@ -0,0 +1,162 @@
+# PR Documentation Update Job - Example Usage
+
+This document provides example usage scenarios for the `pr_docs_update` job.
+
+## Overview
+
+The `pr_docs_update` job automatically updates documentation by analyzing recently merged pull requests. It's designed to:
+- Keep AGENTS.md files up to date with code changes
+- Update copilot-specific instructions when relevant
+- Track processed PRs to avoid duplicates
+- Enable incremental processing across multiple runs
+
+## Basic Usage
+
+### First Run - Process Last 10 PRs
+
+```bash
+# Run all steps to process the 10 most recent merged PRs
+/pr_docs_update.fetch_prs
+/pr_docs_update.analyze_prs
+/pr_docs_update.update_docs
+```
+
+The job will:
+1. Fetch the 10 most recent merged PRs
+2. Filter out any already processed (from state file)
+3. Analyze each PR for documentation relevance
+4. Update appropriate documentation files
+5. Save processed PR numbers to state file
+
+### Subsequent Runs - Process More PRs
+
+```bash
+# Run with a higher count to go deeper into history
+# The state file automatically skips already-processed PRs
+/pr_docs_update.fetch_prs pr_count=20
+/pr_docs_update.analyze_prs
+/pr_docs_update.update_docs
+```
+
+## Advanced Usage
+
+### Custom State File Location
+
+If you want to track state separately for different purposes:
+
+```bash
+/pr_docs_update.fetch_prs state_file=".deepwork/pr_docs_copilot_state.json"
+/pr_docs_update.analyze_prs
+/pr_docs_update.update_docs state_file=".deepwork/pr_docs_copilot_state.json"
+```
+
+### Reprocessing Specific PRs
+
+To reprocess specific PRs, edit the state file:
+
+```bash
+# View current state
+cat .deepwork/pr_docs_state.json
+
+# Remove specific PR numbers you want to reprocess
+# Edit the file and remove those numbers from the "processed_prs" array
+```
+
+### Processing Large Batches
+
+For initial setup or catching up after a long time:
+
+```bash
+# Process 50 PRs at once
+/pr_docs_update.fetch_prs pr_count=50
+/pr_docs_update.analyze_prs
+/pr_docs_update.update_docs
+```
+
+## State File Structure
+
+The state file (`.deepwork/pr_docs_state.json`) tracks processed PRs:
+
+```json
+{
+  "processed_prs": [123, 456, 789],
+  "last_update": "2026-01-19T07:00:00Z",
+  "total_processed": 3
+}
+```
+
+## What Gets Updated?
+
+### AGENTS.md Files
+
+For general code changes that all agents should know about:
+- New jobs or workflows
+- Codebase structure changes
+- New conventions or patterns
+- Important context for working with the code
+
+Located at: `.deepwork/jobs/[job_name]/AGENTS.md`
+
+### Copilot-Specific Instructions
+
+For changes specific to GitHub Copilot integration:
+- Copilot-specific features
+- Copilot workflow changes
+- Copilot tool usage patterns
+
+Located at: `.github/copilot/` or similar copilot instruction locations
+
+## Scheduling Recommendations
+
+- **Weekly**: Run with default settings (10 PRs) to stay current
+- **After Major Releases**: Run with higher count (30-50 PRs)
+- **Initial Setup**: Process all recent PRs in batches
+
+## Integration with Existing Workflows
+
+This job can be combined with other DeepWork jobs:
+
+```bash
+# Update docs from PRs, then commit the changes
+/pr_docs_update.fetch_prs
+/pr_docs_update.analyze_prs
+/pr_docs_update.update_docs
+
+# Review and commit
+/commit.test
+/commit.lint
+/commit.commit_and_push
+```
+
+## Troubleshooting
+
+### No New PRs Found
+
+If all fetched PRs have been processed:
+- Increase the `pr_count` parameter
+- Check the state file to see what's been processed
+- Consider clearing the state file to reprocess all PRs
+
+### GitHub API Rate Limiting
+
+If you hit rate limits:
+- Reduce the `pr_count` parameter
+- Wait for the rate limit to reset
+- Use authenticated GitHub API access if available
+
+### Documentation Files Not Found
+
+If expected documentation files don't exist:
+- The job will notify you about missing files
+- Create the necessary AGENTS.md or instruction files
+- Re-run the update step
+
+## Example Output
+
+After running the job, you'll have:
+
+1. **pr_list.json**: List of fetched PRs (filtered by state)
+2. **pr_analysis.md**: Analysis of each PR with recommendations
+3. **updated_files_list.md**: Summary of what was updated
+4. **Updated documentation files**: AGENTS.md or copilot instructions with new content
+5. **Updated state file**: Tracking newly processed PRs
diff --git a/src/deepwork/standard_jobs/pr_docs_update/job.yml b/src/deepwork/standard_jobs/pr_docs_update/job.yml
new file mode 100644
index 00000000..d057b4da
--- /dev/null
+++ b/src/deepwork/standard_jobs/pr_docs_update/job.yml
@@ -0,0 +1,68 @@
+name: pr_docs_update
+version: "1.0.0"
+summary: "Update documentation from recently merged PRs"
+description: |
+  A workflow that analyzes recently merged pull requests and updates either AGENTS.md
+  or copilot-specific instructions based on the PR content. Tracks processed PRs to
+  avoid duplicates and can be run multiple times to go further back in PR history.
+
+  This job helps maintain up-to-date documentation by:
+  1. Fetching recently merged PRs from GitHub
+  2. Analyzing PR content to identify documentation-relevant changes
+  3. Updating AGENTS.md (for general context) or copilot instructions (for copilot-specific changes)
+  4. Tracking which PRs have been processed to enable incremental updates
+
+  Use this job periodically to keep documentation in sync with code changes.
+
+changelog:
+  - version: "1.0.0"
+    changes: "Initial job creation"
+
+steps:
+  - id: fetch_prs
+    name: "Fetch Recently Merged PRs"
+    description: "Retrieve list of recently merged pull requests from GitHub"
+    instructions_file: steps/fetch_prs.md
+    inputs:
+      - name: pr_count
+        description: "Number of recent PRs to fetch (default: 10)"
+      - name: state_file
+        description: "Path to state file tracking processed PRs (default: .deepwork/pr_docs_state.json)"
+    outputs:
+      - pr_list.json
+    dependencies: []
+
+  - id: analyze_prs
+    name: "Analyze PR Content"
+    description: "Analyze PR content to identify documentation-relevant changes"
+    instructions_file: steps/analyze_prs.md
+    inputs:
+      - file: pr_list.json
+        from_step: fetch_prs
+    outputs:
+      - pr_analysis.md
+    dependencies:
+      - fetch_prs
+
+  - id: update_docs
+    name: "Update Documentation"
+    description: "Update AGENTS.md or copilot instructions based on PR analysis"
+    instructions_file: steps/update_docs.md
+    inputs:
+      - file: pr_analysis.md
+        from_step: analyze_prs
+      - name: state_file
+        description: "Path to state file tracking processed PRs (default: .deepwork/pr_docs_state.json)"
+    outputs:
+      - updated_files_list.md
+    dependencies:
+      - analyze_prs
+    hooks:
+      after_agent:
+        - prompt: |
+            Verify the documentation update is complete:
+            1. Relevant documentation files (AGENTS.md or copilot instructions) were identified
+            2. Documentation was updated with insights from analyzed PRs
+            3. State file was updated with processed PR numbers
+            4. Changes are ready to be committed
+            If ALL criteria are met, include `<promise>✓ Quality Criteria Met</promise>`.
diff --git a/src/deepwork/standard_jobs/pr_docs_update/steps/analyze_prs.md b/src/deepwork/standard_jobs/pr_docs_update/steps/analyze_prs.md
new file mode 100644
index 00000000..84ac31d0
--- /dev/null
+++ b/src/deepwork/standard_jobs/pr_docs_update/steps/analyze_prs.md
@@ -0,0 +1,137 @@
+# Analyze PR Content
+
+## Objective
+
+Analyze the content of unprocessed pull requests to identify which ones contain changes relevant to documentation updates, specifically for AGENTS.md files or copilot-specific instructions.
+
+## Task
+
+Review each PR from the input list and determine what documentation updates are needed based on the PR's changes, title, description, and files modified.
+
+### Step 1: Load PR List
+
+1. **Read the input file** `pr_list.json`
+   - Parse the JSON structure
+   - Extract the list of PRs to analyze
+
+2. **Verify the data**
+   - Ensure PRs have all required fields
+   - Note the repository context
+
+### Step 2: Analyze Each PR
+
+For each PR in the list, evaluate its relevance to documentation:
+
+1. **Examine Changed Files**
+   - Look for patterns that suggest documentation impact:
+     - New features or capabilities added
+     - Changes to job definitions or workflows
+     - API or interface modifications
+     - New standard jobs or tools
+     - Configuration changes
+     - Agent or LLM interaction changes
+   
+2. **Review PR Title and Description**
+   - Identify the purpose of the PR
+   - Note any documentation mentions
+   - Look for keywords like: "add", "new", "feature", "job", "workflow", "agent", "copilot"
+
+3. **Categorize the PR**
+   - **General documentation**: Changes that affect all agents (→ AGENTS.md)
+   - **Copilot-specific**: Changes specific to GitHub Copilot integration (→ copilot instructions)
+   - **Not relevant**: Changes with no documentation impact
+   - **Unclear**: Needs further investigation
+
+4. **Extract Documentation Insights**
+   For relevant PRs, identify:
+   - What was added or changed?
+   - What should agents know about this change?
+   - Which documentation file should be updated?
+   - What specific information should be added?
+
+### Step 3: Determine Documentation Updates
+
+For each relevant PR, create a documentation recommendation:
+
+1. **For AGENTS.md updates**:
+   - New job or workflow information
+   - Codebase structure changes
+   - New conventions or patterns
+   - Important context for all agents
+
+2. **For copilot-specific instructions**:
+   - Copilot-specific features or integrations
+   - Copilot workflow changes
+   - Copilot tool usage patterns
+
+3. **Documentation entry format**:
+   - Clear, concise description
+   - Reference to the PR number
+   - Suggested location in the documentation
+   - Specific text to add or update
+
+### Step 4: Create Analysis Report
+
+Create `pr_analysis.md` with the following structure:
+
+```markdown
+# PR Documentation Analysis
+
+**Repository**: owner/repo-name
+**Analysis Date**: 2026-01-19
+**PRs Analyzed**: 5
+**Documentation Updates Needed**: 3
+
+## PRs Requiring Documentation Updates
+
+### PR #123: Add new feature
+- **Merge Date**: 2026-01-15
+- **Author**: username
+- **URL**: https://github.com/owner/repo/pull/123
+- **Category**: General Documentation (AGENTS.md)
+- **Summary**: This PR adds a new job type for automated testing
+- **Documentation Update**: 
+  - File: `.deepwork/jobs/[job_name]/AGENTS.md`
+  - Section: Job Structure
+  - Content: Document the new automated testing job available in the repository
+  - Rationale: Agents should be aware of this capability for future task planning
+
+### PR #456: Update copilot integration
+- **Merge Date**: 2026-01-14
+- **Author**: username2
+- **URL**: https://github.com/owner/repo/pull/456
+- **Category**: Copilot-Specific Instructions
+- **Summary**: Enhanced copilot workflow with new commands
+- **Documentation Update**: 
+  - File: Copilot-specific instruction files
+  - Section: Available Commands
+  - Content: Document the new `/copilot-command` syntax and usage
+  - Rationale: Copilot agents need to know about new command capabilities
+
+## PRs Not Requiring Updates
+
+### PR #789: Fix typo in README
+- **Reason**: Minor documentation fix, no agent-relevant changes
+```
+
+## Output Format
+
+- **File**: `pr_analysis.md`
+- **Content**: Markdown document with analysis results and recommendations
+- **Location**: Current working directory
+
+## Quality Criteria
+
+- ✓ All PRs from input list analyzed
+- ✓ Each PR categorized (general docs, copilot-specific, not relevant, or unclear)
+- ✓ Documentation updates clearly specified with target file and content
+- ✓ Rationale provided for each recommended update
+- ✓ Summary statistics included (PRs analyzed, updates needed)
+- ✓ Clear, actionable recommendations for the next step
+
+## Notes
+
+- Focus on changes that provide value to AI agents working with the codebase
+- Avoid documenting trivial changes (typos, formatting, etc.)
+- When in doubt, err on the side of including information that might be helpful
+- Consider the perspective of an agent encountering this codebase for the first time
diff --git a/src/deepwork/standard_jobs/pr_docs_update/steps/fetch_prs.md b/src/deepwork/standard_jobs/pr_docs_update/steps/fetch_prs.md
new file mode 100644
index 00000000..1c53ee8e
--- /dev/null
+++ b/src/deepwork/standard_jobs/pr_docs_update/steps/fetch_prs.md
@@ -0,0 +1,108 @@
+# Fetch Recently Merged PRs
+
+## Objective
+
+Retrieve a list of recently merged pull requests from GitHub and filter out any that have already been processed, based on a state file.
+
+## Task
+
+Fetch recently merged PRs from the current repository and prepare a list of unprocessed PRs for analysis.
+
+### Step 1: Understand the Parameters
+
+1. **PR Count**: Determine how many recent PRs to fetch
+   - Default: 10 PRs
+   - If user specified a count, use that value
+   - Can fetch more to go deeper into history
+
+2. **State File**: Load the state file to track processed PRs
+   - Default location: `.deepwork/pr_docs_state.json`
+   - If file doesn't exist, create it with an empty list
+   - File format: `{"processed_prs": [123, 456, 789]}`
+
+### Step 2: Fetch Merged PRs from GitHub
+
+Use GitHub tools available to you to fetch recently merged pull requests:
+
+1. **Get the repository information**
+   - Repository owner and name from current git repository
+   - Use `git remote -v` to identify the GitHub repository
+
+2. **Fetch merged PRs**
+   - Use GitHub API or available GitHub tools to list merged pull requests
+   - Sort by merge date (most recent first)
+   - Fetch the number requested (default 10, or user-specified count)
+   
+3. **For each PR, collect**:
+   - PR number
+   - Title
+   - Description/body
+   - Merge date
+   - Author
+   - Labels (if any)
+   - Changed files (summary)
+   - Link to the PR
+
+### Step 3: Filter Out Already Processed PRs
+
+1. **Load the state file**
+   - Read `.deepwork/pr_docs_state.json` (or user-specified path)
+   - Parse the JSON to get the list of processed PR numbers
+   - If file doesn't exist, treat it as empty (no processed PRs)
+
+2. **Filter the PR list**
+   - Remove PRs whose numbers are in the `processed_prs` list
+   - Keep only unprocessed PRs
+
+3. **Handle the case of no new PRs**
+   - If all fetched PRs have been processed, inform the user
+   - Suggest fetching more PRs by increasing the count
+   - Recommend using a larger `pr_count` value to go deeper into history
+
+### Step 4: Create Output File
+
+Create `pr_list.json` with the following structure:
+
+```json
+{
+  "repository": "owner/repo-name",
+  "fetch_date": "2026-01-19T07:00:00Z",
+  "total_fetched": 10,
+  "new_prs": 5,
+  "prs": [
+    {
+      "number": 123,
+      "title": "Add new feature",
+      "description": "This PR adds...",
+      "merge_date": "2026-01-15T10:30:00Z",
+      "author": "username",
+      "labels": ["enhancement", "documentation"],
+      "changed_files": ["src/feature.py", "docs/README.md"],
+      "url": "https://github.com/owner/repo/pull/123"
+    }
+  ]
+}
+```
+
+## Output Format
+
+- **File**: `pr_list.json`
+- **Content**: JSON structure with repository info and list of unprocessed merged PRs
+- **Location**: Current working directory
+
+## Quality Criteria
+
+- ✓ Repository information correctly identified from git remote
+- ✓ At least the requested number of PRs fetched (unless fewer exist)
+- ✓ State file loaded (or created if missing)
+- ✓ Already-processed PRs filtered out
+- ✓ PR data includes all required fields (number, title, description, merge_date, author, url)
+- ✓ Output file created with valid JSON structure
+- ✓ If no new PRs found, user informed about options to fetch more
+
+## Notes
+
+- This step uses GitHub API access, which may require authentication
+- If you encounter rate limiting, inform the user
+- The state file prevents duplicate processing across multiple runs
+- Users can delete or edit the state file to reprocess specific PRs if needed
diff --git a/src/deepwork/standard_jobs/pr_docs_update/steps/update_docs.md b/src/deepwork/standard_jobs/pr_docs_update/steps/update_docs.md
new file mode 100644
index 00000000..f3ba3a1e
--- /dev/null
+++ b/src/deepwork/standard_jobs/pr_docs_update/steps/update_docs.md
@@ -0,0 +1,168 @@
+# Update Documentation
+
+## Objective
+
+Update AGENTS.md files or copilot-specific instructions based on the PR analysis, and track processed PRs in the state file to prevent duplicate processing in future runs.
+
+## Task
+
+Apply the documentation recommendations from the analysis step, update the appropriate files, and maintain the state file for tracking.
+
+### Step 1: Load the Analysis
+
+1. **Read `pr_analysis.md`**
+   - Parse the recommendations
+   - Identify which files need updates
+   - Group updates by target file
+
+2. **Load the state file**
+   - Read `.deepwork/pr_docs_state.json` (or user-specified path)
+   - Prepare to add newly processed PR numbers
+
+### Step 2: Identify Target Documentation Files
+
+Based on the analysis, determine which documentation files to update:
+
+1. **AGENTS.md files**
+   - Check for existing AGENTS.md files in relevant job directories
+   - Typical locations:
+     - `.deepwork/jobs/[job_name]/AGENTS.md`
+     - `src/deepwork/standard_jobs/[job_name]/AGENTS.md`
+   - If no AGENTS.md exists for relevant jobs, consider creating one
+
+2. **Copilot-specific instructions**
+   - Look for copilot instruction files
+   - Common locations might include:
+     - `.github/copilot/`
+     - `.copilot/`
+     - Other copilot-specific configuration locations
+   - If no copilot instructions exist and updates are needed, note this for the user
+
+3. **Create an update plan**
+   - List all files to be updated
+   - Organize updates by file to minimize context switching
+
+### Step 3: Update Documentation Files
+
+For each documentation file identified:
+
+1. **Review existing content**
+   - Read the current file content
+   - Understand the existing structure and sections
+   - Identify where new content should be added
+
+2. **Apply updates**
+   - Add new information based on PR analysis
+   - Include PR reference numbers for traceability
+   - Follow the existing format and style
+   - Maintain chronological order if applicable
+   - Keep entries clear and concise
+
+3. **AGENTS.md format example**:
+   ```markdown
+   ## Last Updated
+   
+   - Date: 2026-01-19
+   - From: PR documentation sync (PRs #123, #456)
+   
+   ## Recent Changes
+   
+   ### PR #123: New Automated Testing Job
+   - A new job type for automated testing is now available
+   - Location: `.deepwork/jobs/automated_test/`
+   - Use this for test automation workflows
+   - See job.yml for full specification
+   ```
+
+4. **Handle multiple file updates**
+   - Update all relevant files identified in the analysis
+   - Maintain consistency across updates
+   - Ensure cross-references are accurate
+
+### Step 4: Update State File
+
+1. **Collect processed PR numbers**
+   - Extract PR numbers from the analysis
+   - Include all PRs (both relevant and not relevant)
+   - This ensures we don't reprocess any PRs
+
+2. **Update the state file**
+   - Load existing `processed_prs` list
+   - Add new PR numbers to the list
+   - Remove duplicates if any
+   - Sort the list (optional, for readability)
+
+3. **Write the updated state file**
+   ```json
+   {
+     "processed_prs": [123, 456, 789, 1011, 1213],
+     "last_update": "2026-01-19T07:00:00Z",
+     "total_processed": 5
+   }
+   ```
+
+### Step 5: Create Summary Report
+
+Create `updated_files_list.md` documenting what was done:
+
+```markdown
+# Documentation Update Summary
+
+**Date**: 2026-01-19
+**PRs Processed**: 5 (PRs #123, #456, #789, #1011, #1213)
+**Files Updated**: 2
+
+## Updated Files
+
+### .deepwork/jobs/deepwork_jobs/AGENTS.md
+- Added information about new automated testing job (PR #123)
+- Updated last modified date
+
+### .github/copilot/instructions.md
+- Documented new copilot command syntax (PR #456)
+- Added usage examples
+
+## State File Updated
+
+- State file: `.deepwork/pr_docs_state.json`
+- New processed PRs added: 5
+- Total PRs tracked: 15
+
+## PRs Processed But Not Documented
+
+- PR #789: Minor typo fix (not relevant)
+- PR #1011: Internal refactoring (no documentation impact)
+- PR #1213: CI configuration (no agent-visible changes)
+
+## Next Steps
+
+To process more PRs from further back in history, run this job again with:
+- Increase `pr_count` parameter (e.g., 20 or 30)
+- The state file will automatically skip already processed PRs
+```
+
+## Output Format
+
+- **File**: `updated_files_list.md`
+- **Content**: Summary of changes made and files updated
+- **Location**: Current working directory
+
+## Quality Criteria
+
+- ✓ All relevant documentation files identified and updated
+- ✓ Updates include PR references for traceability
+- ✓ Existing file structure and format maintained
+- ✓ State file updated with all processed PR numbers
+- ✓ State file includes metadata (last_update, total_processed)
+- ✓ Summary report created with clear list of changes
+- ✓ Changes are ready to be committed
+- ✓ If no files exist for updates, user is informed
+
+## Notes
+
+- If AGENTS.md files don't exist in expected locations, ask the user if they should be created
+- Maintain the existing tone and style of documentation files
+- Include enough context so future readers understand the changes
+- The state file enables incremental processing across multiple runs
+- Users can manually edit the state file to reprocess specific PRs if needed
+- Consider running a git diff to show the user what changed before committing