diff --git a/.github/workflows/pr-review.yml b/.github/workflows/pr-review.yml
index f7b1e5913..70df5b5ef 100644
--- a/.github/workflows/pr-review.yml
+++ b/.github/workflows/pr-review.yml
@@ -50,7 +50,7 @@ jobs:
       # signs the cascade's escalation comments, and is filtered out of the
       # candidate queue (it can't approve its own PRs). Human reviewers are
       # routed via CODEOWNERS at escalation time, not pinned via env.
-      BOT_USER: ${{ vars.BOT_USER || 'don-petry-bot' }}
+      BOT_USER: ${{ vars.BOT_USER || 'donpetry-bot' }}
       TARGET_ORG: ${{ vars.TARGET_ORG || 'petry-projects' }}
       AGENT_REPO: ${{ github.repository }}
       # Review engine: "claude" or "copilot". Controls which CLI and models are used.
@@ -58,6 +58,8 @@ jobs:
       CLAUDE_CODE_OAUTH_TOKEN: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
       # Copilot engine reuses GH_PAT (a separate user PAT with Copilot subscription).
       COPILOT_GITHUB_TOKEN: ${{ secrets.GH_PAT }}
+      # Gemini engine auth — requires GOOGLE_API_KEY repo secret.
+      GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }}
       # Default to dry-run. workflow_dispatch can override; repository_dispatch
       # (mention) always runs live; vars.LIVE_MODE=true switches scheduled runs live.
       DRY_RUN: ${{ github.event_name == 'repository_dispatch' && 'false' || inputs.dry_run || (vars.LIVE_MODE == 'true' && 'false' || 'true') }}
@@ -107,13 +109,23 @@ jobs:
           case "$REVIEW_ENGINE" in
             claude)
               install_claude
+              # Best-effort install of fallbacks for duck/fallback support
+              npm install -g @google/gemini-cli || true
               if ! env GH_TOKEN="$COPILOT_GITHUB_TOKEN" gh copilot --version > /dev/null 2>&1; then
                 echo "::warning::gh copilot built-in unavailable with user token — rate-limit fallback will be unavailable."
               fi
               ;;
+            gemini)
+              npm install -g @google/gemini-cli
+              install_claude || true
+              ;;
             copilot)
-              gh extension install github/gh-copilot
+              # Best-effort install of copilot if built-in is missing
+              if ! gh copilot --version > /dev/null 2>&1; then
+                gh extension install github/gh-copilot || true
+              fi
               install_claude || true
+              npm install -g @google/gemini-cli || true
               ;;
             *)
               echo "::error::Unknown REVIEW_ENGINE=$REVIEW_ENGINE"
diff --git a/README.md b/README.md
index 1ab9219ca..255364572 100644
--- a/README.md
+++ b/README.md
@@ -46,6 +46,6 @@ Mention trigger: Comment `@petry-review-bot` on any org PR for immediate review.
 
 ## Documentation
 
-- [Architecture & Capabilities](AGENT.md)
-- [Setup Guide](SETUP.md)
-- [Machine User Setup](MACHINE_USER_SETUP.md)
+- [Architecture & Capabilities](docs/pr-review-agent/pr-review-agent.md)
+- [Setup Guide](docs/pr-review-agent/setup.md)
+- [Machine User Setup](docs/pr-review-agent/machine-user-setup.md)
diff --git a/BOT_SETUP.md b/docs/pr-review-agent/bot-setup.md
similarity index 100%
rename from BOT_SETUP.md
rename to docs/pr-review-agent/bot-setup.md
diff --git a/DOCUMENTATION.md b/docs/pr-review-agent/documentation-index.md
similarity index 54%
rename from DOCUMENTATION.md
rename to docs/pr-review-agent/documentation-index.md
index 566041640..5ac1967eb 100644
--- a/DOCUMENTATION.md
+++ b/docs/pr-review-agent/documentation-index.md
@@ -3,20 +3,20 @@
 This repository contains several documentation files describing the PR Review Agent system. Start here to understand which document you need.
 
 ## Quick Start
-- **[SETUP.md](SETUP.md)** ⭐ Start here!
+- **[setup.md](setup.md)** ⭐ Start here!
   - Quick reference for configuration and usage
   - Lists required secrets and variables
   - Shows how to run manually or check logs
   - Includes troubleshooting for common issues
 
 ## Understanding the System
-- **[AGENT.md](AGENT.md)** — Full system documentation
+- **[pr-review-agent.md](pr-review-agent.md)** — Full system documentation
   - Architecture and design philosophy
   - How PR reviews are performed
   - Agent capabilities and limitations
   - Configuration options
 
-- **[IMPLEMENTATION.md](IMPLEMENTATION.md)** — Technical deep dive
+- **[implementation.md](implementation.md)** — Technical deep dive
   - Authentication: machine user with PAT
   - How PR enumeration works
   - Review pipeline architecture
@@ -25,7 +25,7 @@ This repository contains several documentation files describing the PR Review Ag
   - Stuck PR cleanup explained
 
 ## Setting Up Authentication
-- **[MACHINE_USER_SETUP.md](MACHINE_USER_SETUP.md)** — Machine user and PAT setup
+- **[machine-user-setup.md](machine-user-setup.md)** — Machine user and PAT setup
   - Create machine user account and org team
   - Configure CODEOWNERS for code owner approvals
   - Generate fine-grained PAT
@@ -35,49 +35,49 @@ This repository contains several documentation files describing the PR Review Ag
 
 ```
 .github/workflows/
-├── pr-review.yml           # Main hourly review workflow
-└── fix-stuck-prs.yml       # Cleanup for stuck PRs
+    ├── pr-review.yml           # Main hourly review workflow
+    └── fix-stuck-prs.yml       # Cleanup for stuck PRs
 
 scripts/
-├── list-prs.sh             # Find candidate PRs
-├── review-one-pr.sh        # Orchestrate review
-├── post-pr-review.sh       # Post approval and auto-merge
-└── fix-stuck-prs.sh        # Fix old stuck PRs
+    ├── list-prs.sh             # Find candidate PRs
+    ├── review-one-pr.sh        # Orchestrate review
+    ├── post-pr-review.sh       # Post approval and auto-merge
+    └── fix-stuck-prs.sh        # Fix old stuck PRs
 
 prompts/
-├── shared.md               # Shared context
-├── triage.md               # Quick assessment
-├── deep-review.md          # Code analysis
-├── security-audit.md       # Vulnerability check
-└── synthesize.md           # Final verdict
-
-docs/
-├── SETUP.md                # Quick start (read first)
-├── AGENT.md                # Full documentation
-├── IMPLEMENTATION.md       # Technical details
-├── MACHINE_USER_SETUP.md   # Machine user and PAT setup
-└── DOCUMENTATION.md        # This file
+    ├── shared.md               # Shared context
+    ├── triage.md               # Quick assessment
+    ├── deep-review.md          # Code analysis
+    ├── security-audit.md       # Vulnerability check
+    └── synthesize.md           # Final verdict
+
+docs/pr-review-agent/
+    ├── setup.md                # Quick start (read first)
+    ├── pr-review-agent.md                # Full documentation
+    ├── implementation.md       # Technical details
+    ├── machine-user-setup.md   # Machine user and PAT setup
+    └── documentation-index.md        # This file
 ```
 
 ## Common Tasks
 
 ### I want to understand what this agent does
-→ Read [AGENT.md](AGENT.md)
+→ Read [pr-review-agent.md](pr-review-agent.md)
 
 ### I need to set up the agent in a new organization
-→ Follow [MACHINE_USER_SETUP.md](MACHINE_USER_SETUP.md)
+→ Follow [machine-user-setup.md](machine-user-setup.md)
 
 ### The agent isn't working, help!
-→ Check [SETUP.md#troubleshooting](SETUP.md#troubleshooting)
+→ Check [setup.md#troubleshooting](setup.md#troubleshooting)
 
 ### I want to understand the architecture
-→ Read [IMPLEMENTATION.md](IMPLEMENTATION.md)
+→ Read [implementation.md](implementation.md)
 
 ### I need to update configuration
-→ See [SETUP.md#repository-variables](SETUP.md#repository-variables)
+→ See [setup.md#repository-variables](setup.md#repository-variables)
 
 ### I want to run a manual review
-→ See [SETUP.md#running-manually](SETUP.md#running-manually)
+→ See [setup.md#running-manually](setup.md#running-manually)
 
 ## Authentication Method
 
@@ -97,9 +97,9 @@ All secrets must be set in the repository (`Settings → Secrets and variables 
 
 | Secret | Purpose |
 |--------|---------|
-| `CLAUDE_CODE_OAUTH_TOKEN` | Claude Code API access |
-| `DON_PETRY_BOT_PETRY_PROJECT_PAT` | Machine user fine-grained PAT |
-| `COPILOT_GITHUB_TOKEN` | Optional fallback engine |
+| `CLAUDE_CODE_OAUTH_TOKEN` | Claude Code authentication |
+| `DON_PETRY_BOT_GH_PAT` | Machine user fine-grained PAT |
+| `GH_PAT` | User PAT with Copilot subscription |
 
 ## Review Workflow
 
@@ -121,17 +121,17 @@ All secrets must be set in the repository (`Settings → Secrets and variables 
 
 ## Support and Troubleshooting
 
-- **Workflow failing to authenticate**: Check [SETUP.md#troubleshooting](SETUP.md#troubleshooting)
-- **Questions about design**: See [IMPLEMENTATION.md](IMPLEMENTATION.md)
-- **Setup instructions**: Follow [MACHINE_USER_SETUP.md](MACHINE_USER_SETUP.md)
-- **Agent capabilities**: Read [AGENT.md](AGENT.md)
+- **Workflow failing to authenticate**: Check [setup.md#troubleshooting](setup.md#troubleshooting)
+- **Questions about design**: See [implementation.md](implementation.md)
+- **Setup instructions**: Follow [machine-user-setup.md](machine-user-setup.md)
+- **Agent capabilities**: Read [pr-review-agent.md](pr-review-agent.md)
 
 ## Document Maintenance
 
 These documents are kept in sync with the actual implementation. When updating the system:
 1. Update the relevant scripts/workflows
 2. Update the corresponding documentation
-3. Ensure SETUP.md reflects current state
-4. Update IMPLEMENTATION.md if architecture changes
+3. Ensure setup.md reflects current state
+4. Update implementation.md if architecture changes
 
-Last updated: April 26, 2026
+Last updated: May 11, 2026
diff --git a/PR_REVIEW_FAILURE_INVESTIGATION.md b/docs/pr-review-agent/failure-investigation.md
similarity index 100%
rename from PR_REVIEW_FAILURE_INVESTIGATION.md
rename to docs/pr-review-agent/failure-investigation.md
diff --git a/PR_REVIEW_AGENT_FAILURE_REPORT.md b/docs/pr-review-agent/failure-report.md
similarity index 100%
rename from PR_REVIEW_AGENT_FAILURE_REPORT.md
rename to docs/pr-review-agent/failure-report.md
diff --git a/IMPLEMENTATION.md b/docs/pr-review-agent/implementation.md
similarity index 86%
rename from IMPLEMENTATION.md
rename to docs/pr-review-agent/implementation.md
index 738d7d8f2..58800f469 100644
--- a/IMPLEMENTATION.md
+++ b/docs/pr-review-agent/implementation.md
@@ -20,13 +20,13 @@ This document describes the current implementation of the PR Review Agent and ke
 
 ### Token Usage
 
-All workflows use the `DON_PETRY_BOT_PETRY_PROJECT_PAT` org secret directly:
+All workflows use the `DON_PETRY_BOT_GH_PAT` org secret directly:
 ```yaml
 env:
-  GH_TOKEN: ${{ secrets.DON_PETRY_BOT_PETRY_PROJECT_PAT }}
+  GH_TOKEN: ${{ secrets.DON_PETRY_BOT_GH_PAT }}
 ```
 
-The PAT is a fine-grained token scoped to the `petry-projects` organization with 90-day expiry. See [MACHINE_USER_SETUP.md](MACHINE_USER_SETUP.md) for rotation instructions.
+The PAT is a fine-grained token scoped to the `petry-projects` organization with 90-day expiry. See [MACHINE_USER_SETUP.md](machine-user-setup.md) for rotation instructions.
 
 ## PR Enumeration
 
@@ -162,17 +162,22 @@ This script runs with machine user PAT credentials that can approve PRs across t
 
 ## Rate Limiting and Fallback
 
-The `pr-review.yml` workflow includes fallback logic:
+The `scripts/review-batch.sh` script includes fallback logic:
 
 ```bash
+# Fallback chain: claude -> gemini -> copilot
 if [ "$rc" -eq 2 ] && [ "${REVIEW_ENGINE:-claude}" = "claude" ]; then
-  echo "Claude rate limit hit — switching to Copilot engine for remaining PRs"
+  export REVIEW_ENGINE=gemini
+  # ... retry this PR with gemini ...
+fi
+
+if [ "$rc" -eq 2 ] && [ "${REVIEW_ENGINE}" = "gemini" ]; then
   export REVIEW_ENGINE=copilot
-  bash scripts/review-one-pr.sh "$pr_url" || rc=$?
+  # ... retry this PR with copilot ...
 fi
 ```
 
-If Claude hits rate limits, the workflow switches to GitHub Copilot for remaining PRs in the batch. This ensures reviews continue even under high load.
+If Claude hits rate limits, the batch switches to Gemini, and then to GitHub Copilot if needed. This ensures reviews continue even under high load.
 
 ## Metrics and Monitoring
 
@@ -180,7 +185,7 @@ Key metrics tracked in workflow logs:
 - **Reviews posted:** Number of PRs actually approved
 - **No-ops skipped:** PRs already reviewed, not re-reviewed
 - **Failures:** PRs that had errors during review
-- **Engine fallbacks:** Times Claude rate limit triggered Copilot fallback
+- **Engine fallbacks**: Cumulative count and specific engines used (e.g., "gemini, copilot")
 
 View recent runs:
 ```bash
@@ -198,7 +203,7 @@ gh run view <run-id> --repo don-petry/pr-review-agent --log
 
 | Variable | Default | Purpose |
 |----------|---------|---------|
-| `REVIEW_ENGINE` | `claude` | Primary review engine |
+| `REVIEW_ENGINE` | `claude` | Primary review engine: `claude`, `gemini`, or `copilot` |
 | `DRY_RUN` | `true` | If true, don't post reviews |
 | `MAX_PRS` | `10` | Max reviews per run |
 | `CANDIDATE_LIMIT` | `100` | Max candidates scanned |
@@ -209,9 +214,10 @@ gh run view <run-id> --repo don-petry/pr-review-agent --log
 
 | Secret | Purpose |
 |--------|---------|
+| `DON_PETRY_BOT_GH_PAT` | Machine user PAT for GitHub API access (BOT_USER) |
+| `GH_PAT` | User PAT with Copilot subscription (fallback source) |
 | `CLAUDE_CODE_OAUTH_TOKEN` | Claude Code authentication |
-| `DON_PETRY_BOT_PETRY_PROJECT_PAT` | Machine user PAT for GitHub API access |
-| `COPILOT_GITHUB_TOKEN` | GitHub Copilot fallback token |
+| `GOOGLE_API_KEY` | Gemini API authentication |
 
 ## File Structure
 
@@ -234,9 +240,9 @@ gh run view <run-id> --repo don-petry/pr-review-agent --log
 │   ├── synthesize.md              # Combine findings, make decision
 │   └── cascade-action.md          # Review coordination (legacy)
 │
-├── SETUP.md                       # Quick start guide
-├── MACHINE_USER_SETUP.md          # Machine user and PAT setup
-├── IMPLEMENTATION.md              # This file
+├── setup.md                       # Quick start guide
+├── machine-user-setup.md          # Machine user and PAT setup
+├── implementation.md              # This file
 └── README.md                       # Overview
 ```
 
diff --git a/MACHINE_USER_SETUP.md b/docs/pr-review-agent/machine-user-setup.md
similarity index 77%
rename from MACHINE_USER_SETUP.md
rename to docs/pr-review-agent/machine-user-setup.md
index a759c1bdd..71483a460 100644
--- a/MACHINE_USER_SETUP.md
+++ b/docs/pr-review-agent/machine-user-setup.md
@@ -1,6 +1,6 @@
 # Machine User Setup for PR Review Agent
 
-This project authenticates via a **machine user account** with a fine-grained PAT. This replaces the previous GitHub App approach, which could not satisfy CODEOWNERS approval requirements (see [issue #27](https://github.com/don-petry/pr-review-agent/issues/27)).
+This project authenticates via a **machine user account** with a fine-grained PAT. This replaces the previous GitHub App approach, which could not satisfy CODEOWNERS approval requirements (see [issue #27](https://github.com/petry-projects/.github-private/issues/27)).
 
 ## Why Machine User?
 
@@ -11,7 +11,7 @@ This project authenticates via a **machine user account** with a fine-grained PA
 
 ## Step 1: Create the Machine User Account
 
-1. Create a new GitHub account (e.g., `petry-pr-bot`) with a shared org email alias
+1. Create a new GitHub account (e.g., `donpetry-bot`) with a shared org email alias
 2. Add the account to the `petry-projects` organization
 3. Create an org team (e.g., `petry-projects/pr-reviewers`) and add the machine user
 
@@ -45,31 +45,31 @@ Or use path-specific rules as needed.
 ## Step 4: Store the PAT as an Org Secret
 
 ```bash
-gh secret set DON_PETRY_BOT_PETRY_PROJECT_PAT --org petry-projects --body "<paste-token>"
+gh secret set DON_PETRY_BOT_GH_PAT --org petry-projects --body "<paste-token>"
 ```
 
 Or store at repo level if preferred:
 
 ```bash
-gh secret set DON_PETRY_BOT_PETRY_PROJECT_PAT --repo don-petry/pr-review-agent --body "<paste-token>"
+gh secret set DON_PETRY_BOT_GH_PAT --repo petry-projects/.github-private --body "<paste-token>"
 ```
 
 ### Verify the secret is set:
 
 ```bash
-gh secret list --repo don-petry/pr-review-agent
+gh secret list --repo petry-projects/.github-private
 ```
 
-Should show `DON_PETRY_BOT_PETRY_PROJECT_PAT` in the list.
+Should show `DON_PETRY_BOT_GH_PAT` in the list.
 
 ## Step 5: Test the Setup
 
 ```bash
 # Test dry-run
-gh workflow run fix-stuck-prs.yml --repo don-petry/pr-review-agent -f dry_run=true
+gh workflow run fix-stuck-prs.yml --repo petry-projects/.github-private -f dry_run=true
 
 # Check the logs
-gh run view <run-number> --repo don-petry/pr-review-agent --log | grep "Logged in"
+gh run view <run-number> --repo petry-projects/.github-private --log | grep "Logged in"
 ```
 
 Should show the machine user account name.
@@ -79,9 +79,9 @@ Should show the machine user account name.
 Once the migration is validated:
 
 ```bash
-gh secret delete APP_ID --repo don-petry/pr-review-agent
-gh secret delete APP_INSTALLATION_ID --repo don-petry/pr-review-agent
-gh secret delete APP_PRIVATE_KEY --repo don-petry/pr-review-agent
+gh secret delete APP_ID --repo petry-projects/.github-private
+gh secret delete APP_INSTALLATION_ID --repo petry-projects/.github-private
+gh secret delete APP_PRIVATE_KEY --repo petry-projects/.github-private
 ```
 
 Optionally uninstall/delete the GitHub App from [org settings](https://github.com/organizations/petry-projects/settings/apps).
@@ -92,7 +92,7 @@ Fine-grained PATs have a configurable expiry (90-day recommended). To rotate:
 
 1. Sign in as the machine user
 2. Generate a new fine-grained PAT with the same scopes
-3. Update the secret: `gh secret set DON_PETRY_BOT_PETRY_PROJECT_PAT --repo don-petry/pr-review-agent --body "<new-token>"`
+3. Update the secret: `gh secret set DON_PETRY_BOT_GH_PAT --repo petry-projects/.github-private --body "<new-token>"`
 4. Revoke the old token
 
 ## Trade-offs vs GitHub App
@@ -110,7 +110,7 @@ Fine-grained PATs have a configurable expiry (90-day recommended). To rotate:
 ### Issue: "Resource not accessible by integration"
 - Verify the PAT has the required scopes
 - Check the machine user has access to the target repos
-- Ensure `DON_PETRY_BOT_PETRY_PROJECT_PAT` secret is set correctly
+- Ensure `DON_PETRY_BOT_GH_PAT` secret is set correctly
 
 ### Issue: Approval doesn't satisfy code owner requirement
 - Verify the machine user is in the team listed in CODEOWNERS
@@ -118,10 +118,10 @@ Fine-grained PATs have a configurable expiry (90-day recommended). To rotate:
 - Ensure branch protection has `require_code_owner_review: true`
 
 ### Issue: Token expired
-- Generate a new fine-grained PAT and update the `DON_PETRY_BOT_PETRY_PROJECT_PAT` secret
+- Generate a new fine-grained PAT and update the `DON_PETRY_BOT_GH_PAT` secret
 
 ## References
 
 - [GitHub: Managing CODEOWNERS](https://docs.github.com/en/repositories/managing-your-repositorys-settings-and-features/customizing-your-repository/about-code-owners)
 - [GitHub: Fine-grained PATs](https://docs.github.com/en/authentication/keeping-your-account-and-data-secure/managing-your-personal-access-tokens#fine-grained-personal-access-tokens)
-- [Issue #27: Why GitHub Apps can't be in CODEOWNERS](https://github.com/don-petry/pr-review-agent/issues/27)
+- [Issue #27: Why GitHub Apps can't be in CODEOWNERS](https://github.com/petry-projects/.github-private/issues/27)
diff --git a/AGENT.md b/docs/pr-review-agent/pr-review-agent.md
similarity index 75%
rename from AGENT.md
rename to docs/pr-review-agent/pr-review-agent.md
index 8ee9f0d6c..59a4a997c 100644
--- a/AGENT.md
+++ b/docs/pr-review-agent/pr-review-agent.md
@@ -4,19 +4,19 @@ A scheduled GitHub Action that reviews open PRs on don-petry's behalf.
 Runs hourly, classifies risk, auto-approves low/medium-risk PRs that pass all
 quality gates, and escalates high-risk or gated PRs for human review.
 
-Supports two LLM engines via the `REVIEW_ENGINE` repo variable: **Claude** (default)
+Supports three LLM engines via the `REVIEW_ENGINE` repo variable: **Claude** (default), **Gemini**,
 and **Copilot**.
 
 ## How it works
 
 1. **Cron** — `.github/workflows/pr-review.yml` runs at `:07` every hour
    (and on `workflow_dispatch`).
-1a. **@mention** — comment `@petry-review-bot` on any PR to trigger an immediate
+1a. **@mention** — comment `@donpetry-bot` on any PR to trigger an immediate
     review, bypassing the hourly schedule. See [Mention-triggered reviews](#mention-triggered-reviews).
 2. **Enumerate** — `scripts/list-prs.sh` queries GitHub for open PRs across
    every repo the PAT can see (the bot's personal account plus `TARGET_ORG`),
    **excluding PRs authored by `BOT_USER`** (the workflow's authenticated
-   identity, default `don-petry-bot`). GitHub's GraphQL API rejects
+   identity, default `donpetry-bot`). GitHub's GraphQL API rejects
    self-approval unconditionally, so self-authored PRs are unreviewable and
    would otherwise starve the queue. Output: one URL per line.
 3. **Per-PR review** — `scripts/review-one-pr.sh` runs a cascading review
@@ -46,8 +46,12 @@ and **Copilot**.
    model tiers. Different model families have different blind spots — running
    both catches issues that either alone would miss.
 
-   The rubber duck is **always the opposite engine**: if `REVIEW_ENGINE=claude`,
-   the duck is Copilot (GPT-5.4), and vice versa. No extra configuration needed.
+   The rubber duck is **always a diverse engine**:
+   - If `REVIEW_ENGINE=claude`, the duck is Copilot (GPT-5.4).
+   - If `REVIEW_ENGINE=gemini`, the duck is Claude (Sonnet 4.6).
+   - If `REVIEW_ENGINE=copilot`, the duck is Claude (Sonnet 4.6).
+
+   No extra configuration needed.
 
    **Graceful degradation:** if the rubber duck fails (missing credentials,
    CLI not installed, timeout), the cascade continues with the primary deep
@@ -55,14 +59,14 @@ and **Copilot**.
 
    ### Engine model mapping
 
-   | Tier | Claude primary | Copilot primary |
-   |---|---|---|
-   | Triage | Haiku 4.5 | GPT-5-mini |
-   | Deep review | Sonnet 4.6 | GPT-5.2 |
-   | Rubber duck | GPT-5.4 (cross) | Sonnet 4.6 (cross) |
-   | Synthesis | Sonnet 4.6 | GPT-5.2 |
-   | Security audit | Opus 4.6 | GPT-5.4 |
-   | Action / single review | Sonnet 4.6 / Opus 4.6 | GPT-5.2 / GPT-5.4 |
+   | Tier | Claude primary | Gemini primary | Copilot primary |
+   |---|---|---|---|
+   | Triage | Haiku 4.5 | Gemini 2.0 Flash | GPT-5-mini |
+   | Deep review | Sonnet 4.6 | Gemini 1.5 Pro | GPT-5.2 |
+   | Rubber duck | GPT-5.4 (cross) | Sonnet 4.6 (cross) | Sonnet 4.6 (cross) |
+   | Synthesis | Sonnet 4.6 | Gemini 1.5 Pro | GPT-5.2 |
+   | Security audit | Opus 4.6 | Gemini 1.5 Pro | GPT-5.4 |
+   | Action / single review | Sonnet 4.6 / Opus 4.6 | Gemini 1.5 Pro | GPT-5.2 / GPT-5.4 |
 
    **Cost profile:**
    - ~80% of PRs: triage + single confirm (2 calls, ~30s)
@@ -114,19 +118,19 @@ same account both opens PRs (via Claude automation) and tries to approve
 them, every approval will silently fail.
 
 The recommended pattern: create a dedicated **reviewer bot account**
-(e.g. `petry-review-bot`) whose token is stored as `GH_PAT`. PRs are
+(e.g. `donpetry-bot`) whose token is stored as `GH_PAT`. PRs are
 authored by `don-petry`; the bot approves them.
 
 ### 1. Create the reviewer bot account
 
 1. Sign out of GitHub (or use a private browser window).
 2. Go to <https://github.com/signup> and create a new account.
-   - **Username:** e.g. `petry-review-bot`
-   - **Email:** a dedicated alias works well (e.g. `you+petry-review-bot@gmail.com`)
+   - **Username:** e.g. `donpetry-bot`
+   - **Email:** a dedicated alias works well (e.g. `you+donpetry-bot@gmail.com`)
 3. Verify the email address.
 4. Sign back in as `don-petry`.
 5. Go to **github.com/organizations/petry-projects/settings/members** →
-   **Invite member** → enter `petry-review-bot` → Role: **Member**.
+   **Invite member** → enter `donpetry-bot` → Role: **Member**.
 6. Accept the invite from the bot account.
 
 ### 2. Create a classic PAT for the bot
@@ -134,7 +138,7 @@ authored by `don-petry`; the bot approves them.
 A classic PAT is required — fine-grained PATs cannot satisfy the GitHub
 rulesets bypass that org-admin approval requires.
 
-1. Sign in as `petry-review-bot`.
+1. Sign in as `donpetry-bot`.
 2. Go to **Settings → Developer settings → Personal access tokens →
    Tokens (classic)** → **Generate new token (classic)**.
 3. Settings:
@@ -145,10 +149,10 @@ rulesets bypass that org-admin approval requires.
 5. Sign back in as `don-petry` and store the token:
 
 ```
-gh secret set GH_PAT --repo don-petry/pr-review-agent
+gh secret set GH_PAT --repo petry-projects/.github-private
 ```
 
-> **Branch protection / rulesets:** add `petry-review-bot` as an allowed
+> **Branch protection / rulesets:** add `donpetry-bot` as an allowed
 > approver on each protected repo. In the repo ruleset or branch protection
 > settings, ensure the bot is not excluded from the reviewer pool.
 
@@ -165,7 +169,15 @@ claude setup-token
 Store as a repo secret:
 
 ```
-gh secret set CLAUDE_CODE_OAUTH_TOKEN --repo don-petry/pr-review-agent
+gh secret set CLAUDE_CODE_OAUTH_TOKEN --repo petry-projects/.github-private
+```
+
+#### Gemini engine
+
+Requires a Google API Key. Store as a repo secret:
+
+```
+gh secret set GOOGLE_API_KEY --repo petry-projects/.github-private
 ```
 
 #### Copilot engine
@@ -173,30 +185,33 @@ gh secret set CLAUDE_CODE_OAUTH_TOKEN --repo don-petry/pr-review-agent
 Create a GitHub PAT with Copilot scope. Store as a repo secret:
 
 ```
-gh secret set COPILOT_GITHUB_TOKEN --repo don-petry/pr-review-agent
+gh secret set GH_PAT --repo petry-projects/.github-private
 ```
 
 ### 4. Choose your engine
 
 ```
+# Use Gemini:
+gh variable set REVIEW_ENGINE --body gemini --repo petry-projects/.github-private
+
 # Use Copilot (GPT models):
-gh variable set REVIEW_ENGINE --body copilot --repo don-petry/pr-review-agent
+gh variable set REVIEW_ENGINE --body copilot --repo petry-projects/.github-private
 
 # Use Claude (default — no variable needed, or set explicitly):
-gh variable set REVIEW_ENGINE --body claude --repo don-petry/pr-review-agent
+gh variable set REVIEW_ENGINE --body claude --repo petry-projects/.github-private
 ```
 
 ### 5. Test with a dry run
 
 ```
-gh workflow run pr-review.yml --repo don-petry/pr-review-agent -f dry_run=true
-gh run watch --repo don-petry/pr-review-agent
+gh workflow run pr-review.yml --repo petry-projects/.github-private -f dry_run=true
+gh run watch --repo petry-projects/.github-private
 ```
 
 To review a single PR ad-hoc:
 
 ```
-gh workflow run pr-review.yml --repo don-petry/pr-review-agent \
+gh workflow run pr-review.yml --repo petry-projects/.github-private \
   -f pr_url=https://github.com/owner/repo/pull/123 \
   -f dry_run=true
 ```
@@ -208,24 +223,24 @@ but never posts reviews, comments, labels, or reviewer requests. To enable
 live mode:
 
 ```
-gh variable set LIVE_MODE --body true --repo don-petry/pr-review-agent
+gh variable set LIVE_MODE --body true --repo petry-projects/.github-private
 ```
 
 To go back to dry-run:
 
 ```
-gh variable delete LIVE_MODE --repo don-petry/pr-review-agent
+gh variable delete LIVE_MODE --repo petry-projects/.github-private
 ```
 
 A specific run can always be forced either way via the `dry_run` workflow input:
 
 ```
-gh workflow run pr-review.yml --repo don-petry/pr-review-agent -f dry_run=false
+gh workflow run pr-review.yml --repo petry-projects/.github-private -f dry_run=false
 ```
 
 ## Tuning
 
-- **Review engine** — `REVIEW_ENGINE` repo variable: `claude` (default) or
+- **Review engine** — `REVIEW_ENGINE` repo variable: `claude` (default), `gemini`, or
   `copilot`. Controls which CLI and model family is used.
 - **Risk rules** — edit `prompts/shared.md` (taxonomy), or the per-tier
   prompts (`prompts/deep-review.md`, `prompts/security-audit.md`).
@@ -234,31 +249,31 @@ gh workflow run pr-review.yml --repo don-petry/pr-review-agent -f dry_run=false
   PRs from a specific org, or to exclude certain repos).
 - **AI delegation** — set `DELEGATION_ORGS` to a comma-separated list of
   GitHub orgs where AI-assisted fix delegation is enabled:
-  `gh variable set DELEGATION_ORGS --body "petry-projects,don-petry" --repo don-petry/pr-review-agent`
+  `gh variable set DELEGATION_ORGS --body "petry-projects,don-petry" --repo petry-projects/.github-private`
 - **Max review cycles** — how many times the agent delegates to AI before
   escalating to human (default 3):
-  `gh variable set MAX_REVIEW_CYCLES --body 5 --repo don-petry/pr-review-agent`
+  `gh variable set MAX_REVIEW_CYCLES --body 5 --repo petry-projects/.github-private`
 - **Models** — change model IDs in `scripts/engine.sh`. The cascade tiers
   map to: triage → deep → audit → action.
 - **Max PRs per run** — defaults to 10 per cron tick to stay within the 60-min
   job timeout. Override:
-  `gh variable set MAX_PRS --body 15 --repo don-petry/pr-review-agent`
+  `gh variable set MAX_PRS --body 15 --repo petry-projects/.github-private`
 
 ## Mention-triggered reviews
 
-Comment `@petry-review-bot` on any PR to start an immediate, on-demand review
+Comment `@donpetry-bot` on any PR to start an immediate, on-demand review
 without waiting for the next scheduled run. The bot posts an acknowledgement
 within seconds and the full cascade result appears in a few minutes.
 
 **How it flows:**
 
 ```
-PR comment "@petry-review-bot please review"
+PR comment "@donpetry-bot please review"
   → petry-projects/.github: pr-review-mention.yml (listens org-wide)
       → validates commenter trust (OWNER/MEMBER/COLLABORATOR only)
       → posts ack comment "I'm on it..."
       → gh workflow run pr-review.yml --field pr_url=<url> --field force_review=true
-          → don-petry/pr-review-agent: pr-review.yml (per-PR concurrency slot)
+          → petry-projects/.github-private: pr-review.yml (per-PR concurrency slot)
               → scripts/review-one-pr.sh (FORCE_REVIEW=true bypasses idempotency)
                   → cascade as normal → posts review
 ```
@@ -275,19 +290,19 @@ PR comment "@petry-review-bot please review"
 1. Copy [`templates/mention-listener.yml`](templates/mention-listener.yml) to
    `petry-projects/.github` as `.github/workflows/pr-review-mention.yml`.
 
-2. Add the `DON_PETRY_BOT_PETRY_PROJECT_PAT` secret to `petry-projects/.github`
+2. Add the `DON_PETRY_BOT_GH_PAT` secret to `petry-projects/.github`
    (org-level secret or repo secret on `.github`). The PAT needs:
    - **Pull requests: write** — to post the ack comment across petry-projects repos
-   - **Contents: write** (scoped to `don-petry/pr-review-agent`) — to send the
+   - **Contents: write** (scoped to `petry-projects/.github-private`) — to send the
      `repository_dispatch` event (does **not** require `Actions: write`)
 
-3. Ensure `petry-review-bot` has at least **Read** collaborator access on
-   `don-petry/pr-review-agent`.
+3. Ensure `donpetry-bot` has at least **Read** collaborator access on
+   `petry-projects/.github-private`.
 
 ## Architecture
 
 ```
-scripts/engine.sh         ← LLM abstraction (claude/copilot dispatch)
+scripts/engine.sh         ← LLM abstraction (claude/gemini/copilot dispatch)
 scripts/review-one-pr.sh  ← Cascade orchestrator (sources engine.sh)
 scripts/list-prs.sh       ← PR enumeration
 
@@ -302,8 +317,9 @@ prompts/shared.md         ← Shared risk taxonomy and decision gates
 ## Cost
 
 Uses the configured engine's billing:
-- **Claude:** Max plan via OAuth token — no per-token API billing.
-- **Copilot:** Included in GitHub Copilot subscription.
+- **Claude**: Max plan via OAuth token — no per-token API billing.
+- **Gemini**: API-based billing via `GOOGLE_API_KEY`.
+- **Copilot**: Included in GitHub Copilot subscription.
 
 GitHub Actions cost is ~720 runs/month (hourly × 30 days). Runs with zero
 candidate PRs finish in ~10s. Each PR reviewed costs ~2-5 min of runner time
diff --git a/SETUP.md b/docs/pr-review-agent/setup.md
similarity index 77%
rename from SETUP.md
rename to docs/pr-review-agent/setup.md
index c6d85f053..1899d9c08 100644
--- a/SETUP.md
+++ b/docs/pr-review-agent/setup.md
@@ -1,12 +1,13 @@
 # PR Review Agent Setup
 
-This repository automates PR reviews for the `petry-projects` organization using Claude Code or GitHub Copilot.
+This repository automates PR reviews for the `petry-projects` organization using
+Claude Code, Google Gemini, or GitHub Copilot.
 
 ## Quick Start
 
 ### Prerequisites
 - GitHub organization: `petry-projects`
-- Machine user account (e.g., `petry-pr-bot`) added to org team in CODEOWNERS
+- Machine user account (e.g., `donpetry-bot`) added to org team in CODEOWNERS
 - Secrets configured in the repository
 
 ### Repository Secrets Required
@@ -15,15 +16,16 @@ Store these in the repository settings (`Settings → Secrets and variables →
 
 | Secret | Description | Source |
 |--------|-------------|--------|
+| `DON_PETRY_BOT_GH_PAT` | Machine user PAT for GitHub API access | Generated from machine user account settings |
+| `GH_PAT` | User PAT with Copilot subscription (for fallback) | GitHub PAT with Copilot scope |
 | `CLAUDE_CODE_OAUTH_TOKEN` | Claude Code authentication token | `claude setup-token` |
-| `DON_PETRY_BOT_PETRY_PROJECT_PAT` | Machine user fine-grained PAT | Generated from machine user account settings |
-| `COPILOT_GITHUB_TOKEN` | GitHub Copilot token (optional, for fallback) | GitHub PAT with Copilot scope |
+| `GOOGLE_API_KEY` | Gemini API authentication token | Google AI Studio / Vertex AI |
 
 ### Repository Variables (Optional)
 
 | Variable | Default | Purpose |
 |----------|---------|---------|
-| `REVIEW_ENGINE` | `claude` | Primary review engine: `claude` or `copilot` |
+| `REVIEW_ENGINE` | `claude` | Primary review engine: `claude`, `gemini`, or `copilot` |
 | `LIVE_MODE` | `false` | If `true`, reviews are posted live; if `false`, dry-run only |
 | `DELEGATION_ORGS` | Empty | Comma-separated orgs where AI can auto-fix: `petry-projects,don-petry` |
 | `MAX_REVIEW_CYCLES` | `3` | Max review iterations before escalating to human |
@@ -60,14 +62,14 @@ When the agent approves a PR:
 ### Dry-run (test without posting)
 ```bash
 gh workflow run pr-review.yml \
-  --repo don-petry/pr-review-agent \
+  --repo petry-projects/.github-private \
   -f dry_run=true
 ```
 
 ### Review a specific PR
 ```bash
 gh workflow run pr-review.yml \
-  --repo don-petry/pr-review-agent \
+  --repo petry-projects/.github-private \
   -f pr_url=https://github.com/petry-projects/ContentTwin/pull/123 \
   -f dry_run=false
 ```
@@ -76,7 +78,7 @@ gh workflow run pr-review.yml \
 If PRs have agent comments but no approval reviews (from an older bug), use:
 ```bash
 gh workflow run fix-stuck-prs.yml \
-  --repo don-petry/pr-review-agent \
+  --repo petry-projects/.github-private \
   -f dry_run=true
 ```
 
@@ -88,13 +90,13 @@ The `pr-review.yml` workflow runs hourly at `:07` to avoid the top-of-hour cron
 
 To check recent runs:
 ```bash
-gh run list --repo don-petry/pr-review-agent -w pr-review.yml -L 5
+gh run list --repo petry-projects/.github-private -w pr-review.yml -L 5
 ```
 
 ## Troubleshooting
 
 ### Workflow fails: "Resource not accessible by integration"
-- Verify the `DON_PETRY_BOT_PETRY_PROJECT_PAT` secret is set and the token hasn't expired
+- Verify the `DON_PETRY_BOT_GH_PAT` secret is set and the token hasn't expired
 - Check the machine user has access to the target repos
 - Ensure the PAT has the required scopes (Contents: read, Pull requests: read/write, Checks: read)
 
@@ -110,9 +112,9 @@ gh run list --repo don-petry/pr-review-agent -w pr-review.yml -L 5
 
 ## Machine User Details
 
-The bot authenticates as a machine user account with a fine-grained PAT stored as the `DON_PETRY_BOT_PETRY_PROJECT_PAT` secret. The machine user is added to an org team listed in CODEOWNERS, so its approvals satisfy code owner review requirements.
+The bot authenticates as a machine user account with a fine-grained PAT stored as the `DON_PETRY_BOT_GH_PAT` secret. The machine user is added to an org team listed in CODEOWNERS, so its approvals satisfy code owner review requirements.
 
-For full setup instructions, see [MACHINE_USER_SETUP.md](MACHINE_USER_SETUP.md).
+For full setup instructions, see [machine-user-setup.md](machine-user-setup.md).
 
 ## Architecture
 
@@ -140,9 +142,9 @@ prompts/
 - **Fine-grained PAT** scoped to only the permissions needed (Contents: read, Pull requests: read/write, Checks: read)
 - **90-day expiry** on PAT — set a calendar reminder to rotate
 - **PAT** stored only in GitHub Secrets, never logged or committed
-- **Rotation** — generate a new PAT, update `DON_PETRY_BOT_PETRY_PROJECT_PAT` secret, revoke old token
+- **Rotation** — generate a new PAT, update `DON_PETRY_BOT_GH_PAT` secret, revoke old token
 
 ## Related Documentation
 
-- [AGENT.md](AGENT.md) — Full agent capabilities and design
-- [MACHINE_USER_SETUP.md](MACHINE_USER_SETUP.md) — Machine user creation, PAT setup, and rotation
+- [pr-review-agent.md](pr-review-agent.md) — Full agent capabilities and design
+- [machine-user-setup.md](machine-user-setup.md) — Machine user creation, PAT setup, and rotation
diff --git a/STATUS.md b/docs/pr-review-agent/status.md
similarity index 75%
rename from STATUS.md
rename to docs/pr-review-agent/status.md
index d37a17287..0cb2ec3ca 100644
--- a/STATUS.md
+++ b/docs/pr-review-agent/status.md
@@ -6,23 +6,23 @@
 ## What Was Done
 
 ### 1. Machine User Authentication ✅
-- Machine user account with fine-grained PAT stored as `DON_PETRY_BOT_PETRY_PROJECT_PAT` org secret
+- Machine user account with fine-grained PAT stored as `DON_PETRY_BOT_GH_PAT` org secret
 - Added to org team listed in CODEOWNERS — approvals satisfy code owner review requirements
 - Configured permissions: Contents (read), Pull requests (read/write), Checks (read)
-- All workflows use `${{ secrets.DON_PETRY_BOT_PETRY_PROJECT_PAT }}` directly (no token generation step needed)
+- All workflows use `${{ secrets.DON_PETRY_BOT_GH_PAT }}` directly (no token generation step needed)
 
-**Why Machine User?** (migrated from GitHub App — see [issue #27](https://github.com/don-petry/pr-review-agent/issues/27))
+**Why Machine User?** (migrated from GitHub App — see [issue #27](https://github.com/petry-projects/.github-private/issues/27))
 - GitHub Apps cannot be listed in CODEOWNERS (platform limitation)
 - Machine user approvals satisfy `require_code_owner_review` branch protection
 - Simpler auth: direct PAT, no JWT generation step
 
 ### 2. Comprehensive Documentation ✅
 Created in repository:
-- **SETUP.md** — Quick reference guide with configuration, secrets, commands, troubleshooting
-- **IMPLEMENTATION.md** — Technical deep dive: architecture, design decisions, stuck PR cleanup
-- **DOCUMENTATION.md** — Index linking all documentation
+- **setup.md** — Quick reference guide with configuration, secrets, commands, troubleshooting
+- **implementation.md** — Technical deep dive: architecture, design decisions, stuck PR cleanup
+- **documentation-index.md** — Index linking all documentation
 - **Updated README.md** — Status badge and quick links
-- **MACHINE_USER_SETUP.md** — Machine user account and PAT setup instructions
+- **machine-user-setup.md** — Machine user account and PAT setup instructions
 
 ### 3. Stuck PR Cleanup ✅ (Approvals Posted)
 Fixed 24 stuck PRs across three repositories:
@@ -83,65 +83,65 @@ This separation makes the system more robust and easier to debug. Agent decision
 ### Machine User over GitHub App
 Previously used: GitHub App (`petry-projects-pr-review-agent[bot]`) with auto-expiring JWT tokens.
 
-**Why we switched:** GitHub Apps cannot be listed in CODEOWNERS files, so repos with `require_code_owner_review: true` were permanently blocked. Machine user accounts can be added to org teams listed in CODEOWNERS, resolving this limitation. See [issue #27](https://github.com/don-petry/pr-review-agent/issues/27).
+**Why we switched:** GitHub Apps cannot be listed in CODEOWNERS files, so repos with `require_code_owner_review: true` were permanently blocked. Machine user accounts can be added to org teams listed in CODEOWNERS, resolving this limitation. See [issue #27](https://github.com/petry-projects/.github-private/issues/27).
 
 ## Configuration
 
 ### Required Secrets
 ```bash
-gh secret set DON_PETRY_BOT_PETRY_PROJECT_PAT --repo don-petry/pr-review-agent --body "<machine-user-pat>"
-gh secret set CLAUDE_CODE_OAUTH_TOKEN --repo don-petry/pr-review-agent
+gh secret set DON_PETRY_BOT_GH_PAT --repo petry-projects/.github-private --body "<machine-user-pat>"
+gh secret set CLAUDE_CODE_OAUTH_TOKEN --repo petry-projects/.github-private
 ```
 
 ### Optional Variables
 ```bash
 # Engine: claude (default) or copilot
-gh variable set REVIEW_ENGINE --repo don-petry/pr-review-agent --body "claude"
+gh variable set REVIEW_ENGINE --repo petry-projects/.github-private --body "claude"
 
 # Live mode: true to post reviews, false for dry-run
-gh variable set LIVE_MODE --repo don-petry/pr-review-agent --body "true"
+gh variable set LIVE_MODE --repo petry-projects/.github-private --body "true"
 
 # Max reviews per run
-gh variable set MAX_PRS --repo don-petry/pr-review-agent --body "10"
+gh variable set MAX_PRS --repo petry-projects/.github-private --body "10"
 ```
 
 ## Usage
 
 ### Check Recent Reviews
 ```bash
-gh run list --repo don-petry/pr-review-agent -w pr-review.yml -L 5
-gh run view <run-id> --repo don-petry/pr-review-agent --log
+gh run list --repo petry-projects/.github-private -w pr-review.yml -L 5
+gh run view <run-id> --repo petry-projects/.github-private --log
 ```
 
 ### Trigger Manual Review
 ```bash
 # Dry-run (no changes)
-gh workflow run pr-review.yml --repo don-petry/pr-review-agent -f dry_run=true
+gh workflow run pr-review.yml --repo petry-projects/.github-private -f dry_run=true
 
 # Review specific PR
-gh workflow run pr-review.yml --repo don-petry/pr-review-agent \
+gh workflow run pr-review.yml --repo petry-projects/.github-private \
   -f pr_url=https://github.com/petry-projects/ContentTwin/pull/123 \
   -f dry_run=false
 
 # Fix stuck PRs
-gh workflow run fix-stuck-prs.yml --repo don-petry/pr-review-agent -f dry_run=false
+gh workflow run fix-stuck-prs.yml --repo petry-projects/.github-private -f dry_run=false
 ```
 
 ## Known Limitations
 
 1. **Auto-merge requires repo settings**: Auto-merge must be enabled in each repo's settings for `gh pr merge --auto` to work.
 
-2. **PAT rotation**: Machine user PATs expire (90-day recommended). Set a calendar reminder to rotate. See [MACHINE_USER_SETUP.md](MACHINE_USER_SETUP.md).
+2. **PAT rotation**: Machine user PATs expire (90-day recommended). Set a calendar reminder to rotate. See [machine-user-setup.md](machine-user-setup.md).
 
 ## Support & Troubleshooting
 
-See **[SETUP.md](SETUP.md)** for:
+See **[setup.md](setup.md)** for:
 - Permission errors and fixes
 - Authentication issues
 - Workflow failures
 - Rate limiting and fallback engines
 
-See **[IMPLEMENTATION.md](IMPLEMENTATION.md)** for:
+See **[implementation.md](implementation.md)** for:
 - Architecture details
 - Design rationale
 - How each component works
@@ -157,5 +157,5 @@ See **[IMPLEMENTATION.md](IMPLEMENTATION.md)** for:
 ---
 
 **Created by:** Claude Code Agent  
-**Repository:** don-petry/pr-review-agent  
+**Repository:** petry-projects/.github-private  
 **Organization:** petry-projects
diff --git a/scripts/engine.sh b/scripts/engine.sh
index 1c6ecc473..8256aebfc 100644
--- a/scripts/engine.sh
+++ b/scripts/engine.sh
@@ -1,6 +1,6 @@
 #!/usr/bin/env bash
 # Engine abstraction layer for LLM invocations.
-# Supports: claude, copilot
+# Supports: claude, gemini, copilot
 #
 # Sourced by review-one-pr.sh — provides:
 #   run_triage <prompt_file>       — no-tool tier (stdout capture)
@@ -41,6 +41,18 @@ case "$REVIEW_ENGINE" in
     DUCK_ENGINE="copilot"
     DUCK_MODEL="gpt-5.4"
     ;;
+  gemini)
+    ENGINE_TRIAGE_MODEL="gemini-2.0-flash"
+    ENGINE_DEEP_MODEL="gemini-1.5-pro"
+    ENGINE_AUDIT_MODEL="gemini-1.5-pro"
+    ENGINE_ACTION_MODEL="gemini-1.5-pro"
+    ENGINE_SINGLE_MODEL="gemini-1.5-pro"
+    ENGINE_LABEL="triage: gemini-2.0-flash → deep: gemini-1.5-pro + duck: sonnet 4.6 → audit: gemini-1.5-pro"
+    ENGINE_SINGLE_LABEL="single-reviewer mode: gemini-1.5-pro"
+    # Cross-engine rubber duck: use Claude for diversity
+    DUCK_ENGINE="claude"
+    DUCK_MODEL="claude-sonnet-4-6"
+    ;;
   copilot)
     ENGINE_TRIAGE_MODEL="gpt-5-mini"
     ENGINE_DEEP_MODEL="gpt-5.2"
@@ -54,7 +66,7 @@ case "$REVIEW_ENGINE" in
     DUCK_MODEL="claude-sonnet-4-6"
     ;;
   *)
-    echo "::error::Unknown REVIEW_ENGINE='$REVIEW_ENGINE' (expected: claude or copilot)"
+    echo "::error::Unknown REVIEW_ENGINE='$REVIEW_ENGINE' (expected: claude, gemini, or copilot)"
     exit 1
     ;;
 esac
@@ -67,11 +79,11 @@ export DUCK_ENGINE DUCK_MODEL
 echo "    engine: $REVIEW_ENGINE ($ENGINE_LABEL)"
 
 # is_rate_limited <text>
-# Returns 0 (true) if the text looks like a Claude or Copilot rate-limit message.
+# Returns 0 (true) if the text looks like a Claude, Gemini, or Copilot rate-limit message.
 # review-one-pr.sh exits with code 2 when this fires so the caller can switch engines.
 is_rate_limited() {
   local text="$1"
-  echo "$text" | grep -qiE "(hit your limit|rate[ -]?limit|resets [0-9]+(am|pm)|usage limit|quota exceeded|too many requests|exceeded.*quota)"
+  echo "$text" | grep -qiE "(hit your limit|rate[ -]?limit|resets [0-9]+(am|pm)|usage limit|quota exceeded|too many requests|exceeded.*quota|([^0-9]|^)429([^0-9]|$)|exhausted)"
 }
 
 # is_transient_failure <exit_code>
@@ -112,6 +124,13 @@ run_triage() {
           --disallowed-tools "Bash,Read,Write,Edit,Grep,Glob,WebFetch,WebSearch,Task,TodoWrite,NotebookEdit" \
           < "$prompt_file" || rc=$?
         ;;
+      gemini)
+        timeout "$TRIAGE_TIMEOUT_SEC" gemini --prompt "" \
+          --model "$ENGINE_TRIAGE_MODEL" \
+          --approval-mode auto_edit \
+          --output-format text \
+          < "$prompt_file" || rc=$?
+        ;;
       copilot)
         # gh copilot is now a built-in; auth via GH_PAT (user token with Copilot subscription).
         # Model selection is not supported by gh copilot suggest — uses Copilot's default.
@@ -155,6 +174,13 @@ run_agentic() {
         --allowed-tools "Bash,Read,Grep,Glob" \
         < "$prompt_file"
       ;;
+    gemini)
+      timeout "$DEEP_TIMEOUT_SEC" gemini --prompt "" \
+        --model "$model" \
+        --approval-mode auto_edit \
+        --output-format text \
+        < "$prompt_file"
+      ;;
     copilot)
       # gh copilot is now a built-in; auth via GH_PAT (user token with Copilot subscription).
       # Model selection is not supported by gh copilot suggest — uses Copilot's default.
@@ -201,7 +227,7 @@ sys.exit(1)
 
 # run_duck <prompt_file> <model>
 # Cross-engine adversarial "rubber duck" review.
-# Always uses the OPPOSITE engine from REVIEW_ENGINE. Output to stdout.
+# Always uses a different model family from REVIEW_ENGINE. Output to stdout.
 # Strips the opposing engine's credentials to prevent cross-engine leakage.
 run_duck() {
   local prompt_file="$1"
@@ -209,6 +235,7 @@ run_duck() {
   case "$DUCK_ENGINE" in
     claude)
       unset COPILOT_GITHUB_TOKEN 2>/dev/null || true
+      unset GOOGLE_API_KEY 2>/dev/null || true
       timeout "$DUCK_TIMEOUT_SEC" claude --print \
         --model "$model" \
         --permission-mode acceptEdits \
@@ -216,8 +243,18 @@ run_duck() {
         --max-turns 25 \
         < "$prompt_file"
       ;;
+    gemini)
+      unset CLAUDE_CODE_OAUTH_TOKEN 2>/dev/null || true
+      unset COPILOT_GITHUB_TOKEN 2>/dev/null || true
+      timeout "$DUCK_TIMEOUT_SEC" gemini --prompt "" \
+        --model "$model" \
+        --approval-mode auto_edit \
+        --output-format text \
+        < "$prompt_file"
+      ;;
     copilot)
       unset CLAUDE_CODE_OAUTH_TOKEN 2>/dev/null || true
+      unset GOOGLE_API_KEY 2>/dev/null || true
       # gh copilot is now a built-in; auth via GH_PAT (user token with Copilot subscription).
       ( export GH_TOKEN="$COPILOT_GITHUB_TOKEN"
         timeout "$DUCK_TIMEOUT_SEC" gh copilot suggest "$(cat "$prompt_file")" --agent shell
diff --git a/scripts/list-prs.sh b/scripts/list-prs.sh
index 9d79d3ece..00775a2ba 100644
--- a/scripts/list-prs.sh
+++ b/scripts/list-prs.sh
@@ -24,7 +24,7 @@ set -euo pipefail
 # Configurable via environment / repo variables. BOT_USER is the GitHub
 # identity the workflow PAT authenticates as — both the queue scope (which
 # repos to scan) and the self-approval filter use it.
-BOT_USER="${BOT_USER:-don-petry-bot}"
+BOT_USER="${BOT_USER:-donpetry-bot}"
 TARGET_ORG="${TARGET_ORG:-petry-projects}"
 
 # Reject BOT_USER values that aren't valid GitHub usernames before
diff --git a/scripts/repair-pr-approvals.sh b/scripts/repair-pr-approvals.sh
index 5ea97261c..b60c3cc8e 100644
--- a/scripts/repair-pr-approvals.sh
+++ b/scripts/repair-pr-approvals.sh
@@ -81,7 +81,7 @@ while IFS= read -r repo; do
   done < <(gh pr list --repo "$repo" --state open --json url,number,comments --limit 100 2>/dev/null | jq -r '.[] | select(.comments | length > 0) | {url, number} | "\(.url)|\(.number)"')
 done < <(
   {
-    gh repo list "${BOT_USER:-don-petry-bot}" --json nameWithOwner --jq '.[].nameWithOwner' 2>/dev/null || true
+    gh repo list "${BOT_USER:-donpetry-bot}" --json nameWithOwner --jq '.[].nameWithOwner' 2>/dev/null || true
     gh repo list "${TARGET_ORG:-petry-projects}" --json nameWithOwner --jq '.[].nameWithOwner' 2>/dev/null || true
   } | sort -u
 )
diff --git a/scripts/review-batch.sh b/scripts/review-batch.sh
index a65cd3b0c..256f9246a 100644
--- a/scripts/review-batch.sh
+++ b/scripts/review-batch.sh
@@ -30,6 +30,7 @@ actual=0
 skipped_noops=0
 failed=0
 engine_fallbacks=0
+fallback_engines=""
 processed=0
 session_aborted=0
 abort_pr=""
@@ -53,18 +54,34 @@ while IFS= read -r pr_url; do
   rc=0
   bash scripts/review-one-pr.sh "$pr_url" || rc=$?
 
-  # Exit code 2 = engine rate-limited. Switch from claude to copilot once,
-  # then retry this PR. All later PRs in the batch use copilot too.
+  # Exit code 2 = engine rate-limited.
+  # Fallback chain: claude -> gemini -> copilot
   if [ "$rc" -eq 2 ] && [ "${REVIEW_ENGINE:-claude}" = "claude" ]; then
-    if ! gh extension list 2>/dev/null | grep -q copilot; then
-      echo "::warning::Copilot fallback engine unavailable (not installed) — skipping $pr_url and continuing batch"
+    if command -v gemini >/dev/null 2>&1 && [ -n "${GOOGLE_API_KEY:-}" ]; then
+      echo "::warning::Claude rate limit hit — switching to Gemini engine for remaining PRs"
+      export REVIEW_ENGINE=gemini
+      engine_fallbacks=$((engine_fallbacks + 1))
+      fallback_engines="${fallback_engines:+$fallback_engines, }gemini"
+      rc=0
+      bash scripts/review-one-pr.sh "$pr_url" || rc=$?
+    else
+      echo "::warning::Claude rate limit hit but Gemini fallback unavailable (CLI not installed or GOOGLE_API_KEY missing) — falling through to Copilot"
+      rc=2
+      export REVIEW_ENGINE=gemini # Set to gemini so the next block catches it
+    fi
+  fi
+
+  if [ "$rc" -eq 2 ] && [ "${REVIEW_ENGINE}" = "gemini" ]; then
+    if ! gh extension list 2>/dev/null | grep -q copilot && ! gh copilot --version > /dev/null 2>&1; then
+      echo "::warning::Copilot fallback engine unavailable — skipping $pr_url and continuing batch"
       failed=$((failed + 1))
       echo "::endgroup::"
       continue
     fi
-    echo "::warning::Claude rate limit hit — switching to Copilot engine for remaining PRs"
+    echo "::warning::Gemini rate limit hit — switching to Copilot engine for remaining PRs"
     export REVIEW_ENGINE=copilot
     engine_fallbacks=$((engine_fallbacks + 1))
+    fallback_engines="${fallback_engines:+$fallback_engines, }copilot"
     rc=0
     bash scripts/review-one-pr.sh "$pr_url" || rc=$?
   fi
@@ -102,7 +119,7 @@ done < "$PRS_FILE"
 
 remaining=$((total_candidates - processed))
 summary="Summary: $actual reviews posted, $skipped_noops no-ops skipped, $failed failures"
-[ "$engine_fallbacks" -gt 0 ] && summary="$summary, $engine_fallbacks engine fallback(s) to copilot"
+[ "$engine_fallbacks" -gt 0 ] && summary="$summary, $engine_fallbacks engine fallback(s) to $fallback_engines"
 summary="$summary (processed $processed/$total_candidates candidates)"
 
 if [ "$session_aborted" -eq 1 ]; then
diff --git a/scripts/review-one-pr.sh b/scripts/review-one-pr.sh
index 5d18c31a0..407bcb0f9 100644
--- a/scripts/review-one-pr.sh
+++ b/scripts/review-one-pr.sh
@@ -6,8 +6,9 @@
 #
 # Env:
 #   GH_TOKEN              — set by the workflow
-#   REVIEW_ENGINE         — "claude" or "copilot" (default: claude)
+#   REVIEW_ENGINE         — "claude", "gemini", or "copilot" (default: claude)
 #   CLAUDE_CODE_OAUTH_TOKEN — (claude engine) set by the workflow
+#   GOOGLE_API_KEY          — (gemini engine) set by the workflow
 #   COPILOT_GITHUB_TOKEN    — (copilot engine) set by the workflow
 #   DRY_RUN               — "true" or "false"
 #
@@ -160,7 +161,7 @@ This PR has been through $REVIEW_CYCLE automated review cycles (cap: $MAX_CYCLES
 
 Please take a look manually, or close this PR if it's no longer needed. Once a human review resolves the situation, remove the \`needs-human-review\` label and the cascade can be re-engaged on the next push.
 
-_Posted by the ${BOT_USER:-don-petry-bot} PR-review cascade._
+_Posted by the ${BOT_USER:-donpetry-bot} PR-review cascade._
 ESCALATION_END
     gh pr comment "$PR_URL" --body-file "$ESCALATION_BODY" || echo "    warn: gh pr comment failed — escalation marker not posted; will retry next cycle"
     gh pr edit "$PR_URL" --add-label needs-human-review 2>/dev/null || true