Add agent invocation and workflow integration sections

Adds two new sections to PR Testing Agent documentation:

1. When to Use This Agent
   - Automatic invocation triggers (PR creation, CI failures, "ready to merge" checks)
   - Manual invocation scenarios (validating testing claims, reviewing PRs)
   - Usage pattern examples

2. Integration with Existing Workflows
   - Relationship to code review process and code-reviewer agent
   - Integration at each workflow stage (pre-commit, pre-push, during CI, pre-merge)
   - Complete workflow example from code change to merge

These sections clarify when and how to invoke the agent, and how it fits into the existing development workflow alongside git hooks, CI, and code review.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <[email protected]>

Author: justin808

.claude/docs/pr-testing-agent.md

@@ -4,6 +4,116 @@
 
 **Core Principle:** Be deeply suspicious of claims that tests passed unless you have concrete evidence. Assume automated tests have gaps. Manual testing is often required.
 
+## When to Use This Agent
+
+### Automatic Invocation (Recommended)
+
+**Invoke this agent automatically when:**
+
+- Creating a PR (before `gh pr create`)
+- Responding to PR review comments about testing
+- CI failures occur and need investigation
+- Asked to verify if a PR is "ready to merge"
+
+### Manual Invocation
+
+**Explicitly request this agent when:**
+
+- Validating testing claims in PR descriptions
+- Reviewing someone else's PR for testing adequacy
+- Unsure what testing is needed for specific changes
+- Need a comprehensive testing checklist
+
+### Usage Pattern
+
+```bash
+# Before creating PR:
+# "Use the PR Testing Agent to validate my testing before I create this PR"
+
+# During PR review:
+# "Use the PR Testing Agent to check if this PR has adequate testing coverage"
+
+# When investigating failures:
+# "Use the PR Testing Agent to help me debug these CI failures"
+```
+
+## Integration with Existing Workflows
+
+### Relationship to Code Review
+
+**This agent complements but does not replace:**
+
+- Standard code review for logic, design, and maintainability
+- The `code-reviewer` agent (focuses on code quality, security)
+- CI automated checks (provides guidance when they fail)
+
+**This agent specializes in:**
+
+- Testing verification and validation
+- Identifying untested code paths
+- Catching silent failures in build/infrastructure
+- Providing testing checklists based on change type
+- Translating CI failures to local reproduction steps
+
+### Integration with Development Workflow
+
+**Pre-commit:**
+
+- Git hooks handle linting/formatting automatically
+- Agent focuses on which tests to run manually
+
+**Pre-push:**
+
+- Agent validates testing claims before PR creation
+- Generates comprehensive testing checklist
+- Identifies environmental testing limitations
+
+**During CI:**
+
+- Agent helps reproduce CI failures locally
+- Distinguishes pre-existing vs. new failures
+- Maps CI failures to local test commands
+
+**Pre-merge:**
+
+- Agent validates all testing completed
+- Ensures manual testing gaps documented
+- Reviews PR description for testing claims
+
+### Workflow Example
+
+```bash
+# 1. Make code changes
+vim lib/react_on_rails/helper.rb
+
+# 2. Run relevant tests locally (agent suggests which)
+bundle exec rspec spec/react_on_rails/helper_spec.rb
+
+# 3. Commit (hooks auto-lint)
+git commit -m "Fix helper method"
+
+# 4. Before pushing, consult agent
+# "What testing do I need before creating a PR for helper.rb changes?"
+# Agent responds with checklist
+
+# 5. Complete testing checklist
+cd spec/dummy
+bin/dev
+# Test in browser...
+
+# 6. Create PR with testing documentation
+git push
+gh pr create  # Include testing summary from agent
+
+# 7. If CI fails, consult agent again
+# "CI is failing on integration tests, help me reproduce locally"
+# Agent provides exact commands
+
+# 8. Before merge, final validation
+# "Is this PR ready to merge from a testing perspective?"
+# Agent reviews what was tested vs. what's still needed
+```
+
 ## Agent Behavior
 
 ### Default Stance: Skeptical
@@ -415,7 +525,7 @@ bundle exec rspec spec/system/integration_spec.rb
 
 **Template for test status comment:**
 
-````markdown
+```markdown
 ## Testing Status
 
 ### ✅ Verified Locally
@@ -429,15 +539,15 @@ bundle exec rspec spec/system/integration_spec.rb
 
 - [ ] **Build scripts** - CRITICAL: Clean install test in full environment
       `bash
-    rm -rf node_modules && yarn install --frozen-lockfile
-    yarn run yalc:publish
-    `
+  rm -rf node_modules && yarn install --frozen-lockfile
+  yarn run yalc:publish
+  `
 - [ ] **Browser testing** - Dummy app visual inspection
       `bash
-    cd spec/dummy && bin/dev
-    # Visit http://localhost:3000/hello_world
-    # Check console for errors
-    `
+      cd spec/dummy && bin/dev
+  # Visit http://localhost:3000/hello_world
+  # Check console for errors
+  `
 
 ### ❌ Cannot Test (Environment Limitation)
 
@@ -449,7 +559,7 @@ bundle exec rspec spec/system/integration_spec.rb
 - Latest workflow: [link]
 - Known pre-existing failures: None (master passing)
 - New failures introduced: None
-````
+```
 
 ### In Commit Messages