diff --git a/.Jules/palette.md b/.Jules/palette.md
deleted file mode 100644
index fbba55899..000000000
--- a/.Jules/palette.md
+++ /dev/null
@@ -1,16 +0,0 @@
-# Palette's Journal
-
-## 2025-11-20 - [Init]
-
-**Learning:** Initialized journal.
-**Action:** Will record critical UX/a11y learnings here.
-
-## 2025-11-20 - [Async Feedback Pattern]
-
-**Learning:** `useActionState` (React 19) provides `isPending` out of the box, making it trivial to add loading states to forms without extra state management.
-**Action:** Always destructure `isPending` from `useActionState` and use it to disable inputs and show loading feedback.
-
-## 2025-05-18 - CopyButton Accessibility
-
-**Learning:** Icon-only buttons with `sr-only` text often forget to update that text when state changes (e.g., "Copy" -> "Copied"). Static accessible names on dynamic buttons confuse screen reader users.
-**Action:** Always verify that `sr-only` text or `aria-label` updates to reflect the current state (success, loading, etc.).
diff --git a/.gemini/skills/creating-skills/SKILL.md b/.gemini/skills/creating-skills/SKILL.md
new file mode 100644
index 000000000..382d62af2
--- /dev/null
+++ b/.gemini/skills/creating-skills/SKILL.md
@@ -0,0 +1,391 @@
+---
+name: Writing Skills
+description: TDD for process documentation - test with subagents before writing, iterate until bulletproof
+version: 1.0.0
+---
+
+# Writing Skills
+
+## Overview
+
+**Writing skills IS Test-Driven Development applied to process documentation.**
+
+**Skills are stored in `.gemini/skills/` (project-local) or `~/.gemini/skills/` (user-global).** You edit skills directly in the project's codebase.
+
+You write test cases (pressure scenarios with subagents), watch them fail (baseline behavior), write the skill (documentation), watch tests pass (agents comply), and refactor (close loopholes).
+
+**Core principle:** If you didn't watch an agent fail without the skill, you don't know if the skill teaches the right thing.
+
+See `skills/testing/test-driven-development` for the fundamental RED-GREEN-REFACTOR cycle. This skill adapts TDD to documentation.
+
+## What is a Skill?
+
+A **skill** is a reference guide for proven techniques, patterns, or tools. Skills help future Agent instances find and apply effective approaches.
+
+**Skills are:** Reusable techniques, patterns, tools, reference guides
+
+**Skills are NOT:** Narratives about how you solved a problem once
+
+## When to Create a Skill
+
+**Create when:**
+
+- Technique wasn't intuitively obvious to you
+- You'd reference this again across projects
+- Pattern applies broadly (not project-specific)
+- Others would benefit
+
+**Don't create for:**
+
+- One-off solutions
+- Standard practices well-documented elsewhere
+- Project-specific conventions (put in `GEMINI.md`)
+
+## When to Use This Skill
+
+**Situations:**
+
+- When you discover a technique, pattern, or tool worth documenting for reuse.
+- When editing existing skills.
+- When asked to modify skill documentation.
+- When you've written a skill and need to verify it works before deploying.
+
+## TDD Mapping for Skills
+
+| TDD Concept             | Skill Creation                                   |
+| ----------------------- | ------------------------------------------------ |
+| **Test case**           | Pressure scenario with subagent                  |
+| **Production code**     | Skill document (SKILL.md)                        |
+| **Test fails (RED)**    | Agent violates rule without skill (baseline)     |
+| **Test passes (GREEN)** | Agent complies with skill present                |
+| **Refactor**            | Close loopholes while maintaining compliance     |
+| **Write test first**    | Run baseline scenario BEFORE writing skill       |
+| **Watch it fail**       | Document exact rationalizations agent uses       |
+| **Minimal code**        | Write skill addressing those specific violations |
+| **Watch it pass**       | Verify agent now complies                        |
+| **Refactor cycle**      | Find new rationalizations → plug → re-verify     |
+
+The entire skill creation process follows RED-GREEN-REFACTOR.
+
+## Skill Types
+
+### Technique
+
+Concrete method with steps to follow (condition-based-waiting, root-cause-tracing)
+
+### Pattern
+
+Way of thinking about problems (flatten-with-flags, test-invariants)
+
+### Reference
+
+API docs, syntax guides, tool documentation
+
+## Directory Structure
+
+**All skills are stored in `.gemini/skills/`:**
+
+```
+.gemini/skills/
+  skill-name/
+    SKILL.md              # Main reference (required)
+    supporting-file.*     # Only if needed
+```
+
+**Flat namespace** - all skills in one searchable location
+
+**Separate files for:**
+
+1. **Heavy reference** (100+ lines) - API docs, comprehensive syntax
+2. **Reusable tools** - Scripts, utilities, templates
+
+**Keep inline:**
+
+- Principles and concepts
+- Code patterns (< 50 lines)
+- Everything else
+
+## SKILL.md Structure
+
+```markdown
+---
+name: Human-Readable Name
+description: One-line summary of what this does (ASO-critical)
+version: 1.0.0
+---
+
+# Skill Name
+
+## Overview
+
+What is this? Core principle in 1-2 sentences.
+
+## When to Use
+
+[Small inline flowchart IF decision non-obvious]
+
+Bullet list with SYMPTOMS and use cases
+When NOT to use
+
+## Core Pattern (for techniques/patterns)
+
+Before/after code comparison
+
+## Quick Reference
+
+Table or bullets for scanning common operations
+
+## Implementation
+
+Inline code for simple patterns
+Link to file for heavy reference or reusable tools
+
+## Common Mistakes
+
+What goes wrong + fixes
+
+## Real-World Impact (optional)
+
+Concrete results
+```
+
+## Agent Search Optimization (ASO)
+
+**Critical for discovery:** Future Agents need to FIND your skill.
+
+### 1. Rich Description
+
+Include SYMPTOMS and key keywords in the `description` field, as this is the primary discovery mechanism:
+
+```yaml
+# ❌ BAD: Too abstract
+description: For async testing
+
+# ✅ GOOD: Symptoms and context
+description: Fix flaky async tests using condition-based waiting. Handles timeouts, race conditions, and CI failures.
+```
+
+### 2. Keyword Coverage
+
+Use words an Agent would search for:
+
+- Error messages: "Hook timed out", "ENOTEMPTY", "race condition"
+- Symptoms: "flaky", "hanging", "zombie", "pollution"
+- Synonyms: "timeout/hang/freeze", "cleanup/teardown/afterEach"
+- Tools: Actual commands, library names, file types
+
+### 3. Descriptive Naming
+
+**Use active voice, verb-first:**
+
+- ✅ `creating-skills` not `skill-creation`
+- ✅ `testing-skills-with-subagents` not `subagent-skill-testing`
+
+### 4. Token Efficiency (Critical)
+
+**Problem:** Skills are loaded into the Agent's context. Every token counts.
+
+**Target word counts:**
+
+- Getting-started workflows: <150 words each
+- Other skills: <500 words (be concise)
+
+**Techniques:**
+
+**Move details to tool help:**
+
+```bash
+# ✅ GOOD: Reference --help
+npm run test -- --help
+```
+
+**Use cross-references:**
+
+```markdown
+# ✅ GOOD: Reference other skill
+
+Always use subagents. See `skills/getting-started` for workflow.
+```
+
+**Eliminate redundancy:**
+
+- Don't repeat what's in cross-referenced skills
+- Don't explain what's obvious from context
+
+**Name by what you DO or core insight:**
+
+- ✅ `condition-based-waiting` > `async-test-helpers`
+- ✅ `using-skills` not `skill-usage`
+- ✅ `flatten-with-flags` > `data-structure-refactoring`
+- ✅ `root-cause-tracing` > `debugging-techniques`
+
+**Gerunds (-ing) work well for processes:**
+
+- `creating-skills`, `testing-skills`, `debugging-with-logs`
+
+### 4. Content Repetition
+
+Mention key concepts in the description, overview, and section headers to increase Grep hits.
+
+### 5. Cross-Referencing Other Skills
+
+Use path format relative to the skills root:
+
+- ✅ Good: `skills/testing/test-driven-development`
+
+## Flowchart Usage
+
+```dot
+digraph when_flowchart {
+    "Need to show information?" [shape=diamond];
+    "Decision where I might go wrong?" [shape=diamond];
+    "Use markdown" [shape=box];
+    "Small inline flowchart" [shape=box];
+
+    "Need to show information?" -> "Decision where I might go wrong?" [label="yes"];
+    "Decision where I might go wrong?" -> "Small inline flowchart" [label="yes"];
+    "Decision where I might go wrong?" -> "Use markdown" [label="no"];
+}
+```
+
+**Use flowcharts ONLY for:**
+
+- Non-obvious decision points
+- Process loops where you might stop too early
+
+See `docs/ui-patterns/graphviz-conventions.dot` for graphviz style rules.
+
+## Code Examples
+
+**One excellent example beats many mediocre ones.**
+
+**Good example:**
+
+- Complete and runnable
+- Well-commented explaining WHY
+- From real scenario
+- Shows pattern clearly
+
+## The Iron Law (Same as TDD)
+
+```
+NO SKILL WITHOUT A FAILING TEST FIRST
+```
+
+This applies to NEW skills AND EDITS to existing skills.
+
+Write skill before testing? Delete it. Start over.
+Edit skill without testing? Same violation.
+
+**No exceptions:**
+
+- Not for "simple additions"
+- Not for "just adding a section"
+- Not for "documentation updates"
+- Delete means delete
+
+See `skills/testing/test-driven-development` for why this matters.
+
+## Testing All Skill Types
+
+### Discipline-Enforcing Skills
+
+**Test with:** Pressure scenarios (time + sunk cost + exhaustion). Identify rationalizations and add counters.
+
+### Technique Skills
+
+**Test with:** Application scenarios. Can the Agent apply the technique to new scenarios?
+
+### Pattern Skills
+
+**Test with:** Recognition scenarios. Does the Agent know when pattern applies?
+
+### Reference Skills
+
+**Test with:** Retrieval scenarios. Can the Agent find and use the right info?
+
+## Common Rationalizations for Skipping Testing
+
+| Excuse                     | Reality                                        |
+| -------------------------- | ---------------------------------------------- |
+| "Skill is obviously clear" | Clear to you ≠ clear to other agents. Test it. |
+| "Testing is overkill"      | Untested skills have issues. Always.           |
+| "I'm confident it's good"  | Overconfidence guarantees issues. Test anyway. |
+
+## Bulletproofing Skills Against Rationalization
+
+### Close Every Loophole Explicitly
+
+Forbid specific workarounds:
+
+```markdown
+Write code before test? Delete it. Start over.
+**No exceptions:** Don't keep it as "reference". Delete means delete.
+```
+
+### Build Rationalization Table
+
+Capture excuses from baseline testing and provide counters.
+
+### Create Red Flags List
+
+```markdown
+## Red Flags - STOP and Start Over
+
+- Code before test
+- "I already manually tested it"
+- "It's about spirit not ritual"
+```
+
+## RED-GREEN-REFACTOR for Skills
+
+### RED: Write Failing Test (Baseline)
+
+Run scenario with subagent WITHOUT the skill. Document baseline behavior and rationalizations verbatim.
+
+### GREEN: Write Minimal Skill
+
+Write skill addressing those specific failures. Verify Agent now complies.
+
+### REFACTOR: Close Loopholes
+
+Add explicit counters for any new rationalizations found during testing.
+
+## STOP: Before Moving to Next Skill
+
+**After writing ANY skill, you MUST STOP and complete the verification process.**
+
+**Do NOT:**
+
+- Create multiple skills in batch without testing each.
+- Skip testing because "batching is more efficient".
+
+## Skill Creation Checklist
+
+- [ ] Create pressure scenarios
+- [ ] Run scenarios WITHOUT skill - document baseline behavior
+- [ ] Name describes what you DO
+- [ ] Frontmatter with rich description (primary discovery)
+- [ ] Address baseline failures
+- [ ] One excellent example
+- [ ] Run scenarios WITH skill - verify compliance
+- [ ] Identify NEW rationalizations and add explicit counters
+- [ ] Re-test until bulletproof
+
+## Discovery Workflow
+
+How future Agents find your skill:
+
+1. **Encounter problem** ("tests are flaky")
+2. **Automatic Discovery** (Agent scans `.gemini/skills/` descriptions)
+3. **Finds SKILL.md** (rich description matches)
+4. **Scans overview** (is this relevant?)
+5. **Reads patterns** (quick reference table)
+
+## The Bottom Line
+
+**Creating skills IS TDD for process documentation.**
+
+Same Iron Law: No skill without failing test first.
+Same cycle: RED (baseline) → GREEN (write skill) → REFACTOR (close loopholes).
+Same benefits: Better quality, fewer surprises, bulletproof results.
diff --git a/.gemini/skills/jules-pr-manager/SKILL.md b/.gemini/skills/jules-pr-manager/SKILL.md
new file mode 100644
index 000000000..10084ddec
--- /dev/null
+++ b/.gemini/skills/jules-pr-manager/SKILL.md
@@ -0,0 +1,119 @@
+---
+name: Jules PR Manager
+description: Manage Google Jules PR lifecycle - batch investigation, reverse-order processing, and tiered decision presentation.
+version: 2.7.0
+---
+
+# Jules PR Manager
+
+## Overview
+
+**Manage the lifecycle of PRs created by Google's Jules agent (Gemini 2.5 Pro).**
+
+This skill orchestrates a **Batch-Analyze-Present-Execute** loop. It strictly prioritizes **finishing work** (merging) before **starting work** (vetting).
+
+⚠️ **MANDATORY**: You MUST use the bash scripts in `.gemini/skills/jules-pr-manager/` for ALL interactions. Do not assemble `gh` commands manually.
+
+## State Machine (The Workflow)
+
+Process PRs in this order (closest to completion first):
+
+### 1. Merge Candidates (The Exit)
+
+- **Definition**: PRs vetted, Copilot Approved, CI Passed, No Conflicts.
+- **Action**: Run `./merge.sh <ID> ["message"]`.
+
+### 2. Copilot Review (`jules:copilot-review`)
+
+- **Condition: Copilot left comments**:
+  - Run `./request-changes.sh <ID> "Refined Copilot comments..."`.
+  - Run `./label.sh <ID> remove jules:copilot-review`.
+  - Run `./label.sh <ID> add jules:changes-requested`.
+- **Condition: CI Failed**:
+  - Run `./request-changes.sh <ID> "Fix CI failures: <details>"`.
+  - Run `./label.sh <ID> remove jules:copilot-review`.
+  - Run `./label.sh <ID> add jules:changes-requested`.
+- **Condition: Approved**: Move to Merge Candidate.
+
+### 3. Transition to Copilot Review
+
+- **Condition**: `jules:vetted` AND CI Passed AND NOT `jules:copilot-review`.
+- **Action**:
+  - Run `./mark-ready.sh <ID>`.
+  - Run `./label.sh <ID> add jules:copilot-review`.
+
+### 4. Changes Requested (`jules:changes-requested`)
+
+- **Condition: Jules Responded (New commits/comments)**:
+  - If fixes are good -> Move to "Transition to Copilot Review".
+  - If fixes are bad -> Run `./request-changes.sh <ID> "Still needs work: <details>"`.
+- **Condition: Stalled (>30m silence)**:
+  - Run `./request-changes.sh <ID> "Ping @jules - status update?"`.
+
+### 5. New PRs (Unvetted)
+
+- **Action 1: Duplicate Detection**:
+  - Run `./diff.sh <ID1> <ID2>` to compare.
+  - Pick the best implementation (the "Keeper").
+  - Run `./request-changes.sh <Keeper_ID> "Incorporate good parts from #<Loser_ID>..."`.
+  - Run `./close.sh <Loser_ID> "Duplicate of #<Keeper_ID>"`.
+- **Action 2: Vetting**:
+  - Review code. If worth doing:
+    - Run `./label.sh <ID> add jules:vetted`.
+    - If CI is already passing, move to "Transition to Copilot Review".
+  - If NOT worth doing:
+    - Run `./close.sh <ID> "Rationale for rejection"`.
+
+---
+
+## Execution Model
+
+### 1. Batch Investigation
+
+- Run `./investigate.sh` to get a consolidated JSON (check `lastInstructionAcknowledged` field to confirm Jules has read feedback). of all Jules PRs and their filtered activity timelines.
+- Log the status and plan for every PR before presenting.
+
+### 2. Decision Collection
+
+- Group actions into: **Merge Decisions**, **Vetting Decisions**, and **Trivial Batch** (label changes, automated request-changes).
+
+### 3. Tiered Presentation
+
+Present decisions to the User in this EXACT order:
+
+1. **Merge Decisions**: One by one.
+2. **Vetting Requests**: One by one (include deep opinion on rationale/impact).
+3. **Trivial Batch**: As a numbered list for bulk approval.
+
+### 4. Batch Execution
+
+- Execute all approved actions in one turn. Use `&` to run script calls in the background where appropriate to improve performance.
+
+---
+
+## Automation Scripts Reference
+
+| Script                                  | Purpose                                | Review Type       | Tags @jules |
+| :-------------------------------------- | :------------------------------------- | :---------------- | :---------- |
+| `./investigate.sh`                      | One-shot fetch of all PRs + timelines. | N/A               | No          |
+| `./view-pr.sh <id>`                     | Full unfiltered JSON for one PR.       | N/A               | No          |
+| `./diff.sh <id1> [id2]`                 | View diff or compare two PRs.          | N/A               | No          |
+| `./request-changes.sh <id> <msg>`       | Demand work/fixes or nudge.            | `REQUEST_CHANGES` | **Yes**     |
+| `./merge.sh <id> [msg]`                 | Finalize and merge.                    | `APPROVE`         | **Yes**     |
+| `./close.sh <id> [msg]`                 | Close the PR.                          | N/A               | No          |
+| `./mark-ready.sh <id>`                  | Trigger Copilot (gh pr ready).         | N/A               | No          |
+| `./label.sh <id> <add\|remove> <label>` | Manage state labels.                   | N/A               | No          |
+
+---
+
+## GitHub Labels Reference
+
+- `jules:vetted`: Approved for processing.
+- `jules:copilot-review`: Waiting for Copilot.
+- `jules:changes-requested`: Waiting for Jules (Fixes/Comments).
+- `jules:agent-stalled`: >30m silence.
+- `jules:merge-conflicts`: Git conflict.
+
+## Agent Search Optimization (ASO)
+
+**Keywords**: jules, pr manager, scripts, automation, batch, merge, vetting, duplicates, tiered presentation
diff --git a/.gemini/skills/jules-pr-manager/close.sh b/.gemini/skills/jules-pr-manager/close.sh
new file mode 100644
index 000000000..bd3f5ad77
--- /dev/null
+++ b/.gemini/skills/jules-pr-manager/close.sh
@@ -0,0 +1,10 @@
+#!/bin/bash
+PR_ID=$1
+shift
+MSG="$*"
+if [ -z "$PR_ID" ]; then echo "Usage: $0 <pr-id> [message]"; exit 1; fi
+if [ -z "$MSG" ]; then
+  gh pr close "$PR_ID"
+else
+  gh pr close "$PR_ID" --comment "$MSG"
+fi
diff --git a/.gemini/skills/jules-pr-manager/diff.sh b/.gemini/skills/jules-pr-manager/diff.sh
new file mode 100644
index 000000000..644343310
--- /dev/null
+++ b/.gemini/skills/jules-pr-manager/diff.sh
@@ -0,0 +1,19 @@
+#!/bin/bash
+ID1=$1
+ID2=$2
+
+if [ -z "$ID1" ]; then
+  echo "Usage: $0 <id1> [id2]"
+  exit 1
+fi
+
+if [ -z "$ID2" ]; then
+  # Single PR diff
+  gh pr diff "$ID1"
+else
+  # Compare two PRs
+  gh pr diff "$ID1" > "$ID1.diff"
+  gh pr diff "$ID2" > "$ID2.diff"
+  diff -u "$ID1.diff" "$ID2.diff"
+  rm "$ID1.diff" "$ID2.diff"
+fi
diff --git a/.gemini/skills/jules-pr-manager/investigate.sh b/.gemini/skills/jules-pr-manager/investigate.sh
new file mode 100644
index 000000000..d182b434d
--- /dev/null
+++ b/.gemini/skills/jules-pr-manager/investigate.sh
@@ -0,0 +1,39 @@
+#!/bin/bash
+# .gemini/skills/jules-pr-manager/investigate.sh
+# Investigates all open Jules PRs and their timelines in a context-efficient way.
+# Now includes verification of Jules acknowledgement (:eyes:).
+
+# 1. Get Repo Info
+REPO_INFO=$(gh repo view --json owner,name)
+OWNER=$(echo "$REPO_INFO" | jq -r .owner.login)
+NAME=$(echo "$REPO_INFO" | jq -r .name)
+
+# 2. List all open Jules PRs
+PRS=$(gh pr list --search "author:app/google-labs-jules is:open" --json number,title,labels,updatedAt,mergeable,statusCheckRollup,isDraft,body)
+
+# 3. For each PR, fetch and filter the timeline
+echo "$PRS" | jq -c '.[]' | while read -r pr; do
+  NUMBER=$(echo "$pr" | jq -r .number)
+  
+  # Fetch timeline
+  TIMELINE=$(gh api "repos/$OWNER/$NAME/issues/$NUMBER/timeline" --paginate)
+  
+  # Filter timeline for efficiency
+  # Note: jq sub() needs double backslashes for literals in shell heredocs/scripts
+  FILTERED_TIMELINE=$(echo "$TIMELINE" | jq -c '[.[] | select(.event | test("committed|reviewed|commented|labeled|unlabeled")) | {
+    event: .event,
+    actor: (.actor.login // .user.login // .author.name),
+    timestamp: (.created_at // .submitted_at // .author.date),
+    body: (.body | select(. != null) | sub("\\n---\\n\\*PR created automatically.*"; "") | sub("^@jules "; "") | .[0:500]),
+    state: .state,
+    label: .label.name,
+    sha: (.sha // .commit_id),
+    acknowledged: (if .reactions then (.reactions.eyes > 0) else false end)
+  }]')
+  
+  # Determine if the LAST review/comment (non-Jules) was acknowledged
+  LAST_INSTRUCTION_ACK=$(echo "$FILTERED_TIMELINE" | jq -r 'map(select(.actor != "google-labs-jules[bot]" and .actor != "vercel[bot]" and .actor != "copilot-pull-request-reviewer" and (.event == "reviewed" or .event == "commented"))) | last | .acknowledged')
+  
+  # Add filtered timeline and acknowledgement status to the PR object
+  echo "$pr" | jq --argjson timeline "$FILTERED_TIMELINE" --arg ack "$LAST_INSTRUCTION_ACK" '. + {timeline: $timeline, lastInstructionAcknowledged: ($ack == "true")}'
+done | jq -s '.'
diff --git a/.gemini/skills/jules-pr-manager/label.sh b/.gemini/skills/jules-pr-manager/label.sh
new file mode 100644
index 000000000..8e60f036d
--- /dev/null
+++ b/.gemini/skills/jules-pr-manager/label.sh
@@ -0,0 +1,16 @@
+#!/bin/bash
+PR_ID=$1
+ACTION=$2
+LABEL=$3
+if [ -z "$PR_ID" ] || [ -z "$ACTION" ] || [ -z "$LABEL" ]; then 
+  echo "Usage: $0 <pr-id> <add|remove> <label>"
+  exit 1
+fi
+if [ "$ACTION" == "add" ]; then
+  gh pr edit "$PR_ID" --add-label "$LABEL"
+elif [ "$ACTION" == "remove" ]; then
+  gh pr edit "$PR_ID" --remove-label "$LABEL"
+else
+  echo "Action must be add or remove"
+  exit 1
+fi
diff --git a/.gemini/skills/jules-pr-manager/mark-ready.sh b/.gemini/skills/jules-pr-manager/mark-ready.sh
new file mode 100644
index 000000000..e6eefb561
--- /dev/null
+++ b/.gemini/skills/jules-pr-manager/mark-ready.sh
@@ -0,0 +1,4 @@
+#!/bin/bash
+PR_ID=$1
+if [ -z "$PR_ID" ]; then echo "PR ID required"; exit 1; fi
+gh pr ready "$PR_ID"
diff --git a/.gemini/skills/jules-pr-manager/merge.sh b/.gemini/skills/jules-pr-manager/merge.sh
new file mode 100644
index 000000000..6d7edef6a
--- /dev/null
+++ b/.gemini/skills/jules-pr-manager/merge.sh
@@ -0,0 +1,20 @@
+#!/bin/bash
+# .gemini/skills/jules-pr-manager/merge.sh
+PR_ID=$1
+shift
+MSG="$*"
+if [ -z "$PR_ID" ]; then echo "Usage: $0 <pr-id> [approval-message]"; exit 1; fi
+
+# Default message
+if [ -z "$MSG" ]; then MSG="Approved. Merging changes."; fi
+
+# Ensure @jules is in the message for the review record
+if [[ ! "$MSG" == *"@jules"* ]]; then
+  MSG="@jules $MSG"
+fi
+
+echo "Posting approval review for #$PR_ID..."
+gh pr review "$PR_ID" --approve --body "$MSG"
+
+echo "Merging #$PR_ID..."
+gh pr merge "$PR_ID" --squash --delete-branch
diff --git a/.gemini/skills/jules-pr-manager/request-changes.sh b/.gemini/skills/jules-pr-manager/request-changes.sh
new file mode 100644
index 000000000..3b846e862
--- /dev/null
+++ b/.gemini/skills/jules-pr-manager/request-changes.sh
@@ -0,0 +1,10 @@
+#!/bin/bash
+PR_ID=$1
+shift
+MSG="$*"
+if [ -z "$PR_ID" ] || [ -z "$MSG" ]; then echo "Usage: $0 <pr-id> <message>"; exit 1; fi
+# Ensure @jules is in the message
+if [[ ! "$MSG" == *"@jules"* ]]; then
+  MSG="@jules $MSG"
+fi
+gh pr review "$PR_ID" --request-changes --body "$MSG"
diff --git a/.gemini/skills/jules-pr-manager/view-pr.sh b/.gemini/skills/jules-pr-manager/view-pr.sh
new file mode 100644
index 000000000..6d07360b5
--- /dev/null
+++ b/.gemini/skills/jules-pr-manager/view-pr.sh
@@ -0,0 +1,4 @@
+#!/bin/bash
+PR_ID=$1
+if [ -z "$PR_ID" ]; then echo "PR ID required"; exit 1; fi
+gh pr view "$PR_ID" --json number,title,body,labels,state,mergeable,statusCheckRollup,reviews,comments,commits
diff --git a/.github-staging/README.md b/.github-staging/README.md
deleted file mode 100644
index 729797a3d..000000000
--- a/.github-staging/README.md
+++ /dev/null
@@ -1,37 +0,0 @@
-# GitHub Configuration Staging
-
-This directory contains GitHub configuration files that require manual review and placement into the proper `.github/` directory.
-
-## Why This Directory Exists
-
-GitHub App security restrictions prevent automated pushing of certain files (especially workflow files) to `.github/workflows/`. This staging area allows files to be version controlled while requiring explicit human review before activation.
-
-## Files Pending Review
-
-- **workflows/** - GitHub Actions CI/CD workflows
-- **dependabot.yml** - Automated dependency updates configuration
-
-## Review Process
-
-1. **Review the file** - Understand what it does, verify it's safe
-2. **Test if possible** - For workflows, consider testing in a fork first
-3. **Move to proper location**:
-
-   ```bash
-   # For workflows
-   cp -r .github-staging/workflows .github/
-   git add .github/workflows
-
-   # For dependabot
-   cp .github-staging/dependabot.yml .github/
-   git add .github/dependabot.yml
-
-   git commit -m "chore: activate GitHub configuration after review"
-   git push
-   ```
-
-## Copilot Instructions
-
-Files in this directory MUST be reviewed before being moved to the actual `.github/` directory. Do not automatically move or activate these files - they require human approval.
-
-See: `.github/copilot-instructions.md` for the review checklist.
diff --git a/.github-staging/workflows/README.md b/.github-staging/workflows/README.md
deleted file mode 100644
index e4c2026c1..000000000
--- a/.github-staging/workflows/README.md
+++ /dev/null
@@ -1,33 +0,0 @@
-# CI Workflow - Manual Addition Required
-
-This directory contains the CI workflow that couldn't be automatically pushed due to GitHub App workflow permissions.
-
-## To Add the CI Workflow
-
-**Option 1: Via GitHub UI**
-
-1. Go to your repository on GitHub
-2. Navigate to `.github/workflows/`
-3. Create a new file: `ci.yml`
-4. Paste the contents from `ci.yml` in this directory
-
-**Option 2: Grant Workflow Permissions**
-
-1. Update GitHub App permissions to allow workflow file modifications
-2. Push the workflow file from your local machine
-
-## CI Workflow Contents
-
-See the `ci.yml` file in this directory for the complete workflow configuration.
-
-The workflow includes:
-
-- TypeScript type checking
-- ESLint linting
-- Prettier format checking
-- Production build verification
-- Unit & integration tests
-- Gitleaks secret scanning
-- npm audit security checks
-
-**Status**: Pending manual addition via GitHub UI
diff --git a/.github-staging/workflows/ci.yml b/.github-staging/workflows/ci.yml
deleted file mode 100644
index ceaf0f328..000000000
--- a/.github-staging/workflows/ci.yml
+++ /dev/null
@@ -1,203 +0,0 @@
-name: CI
-
-on:
-  push:
-    branches: ["main", "claude/**"]
-  pull_request:
-    branches: ["main"]
-
-jobs:
-  typecheck:
-    name: TypeScript Type Check
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@v4
-
-      - name: Setup Node.js
-        uses: actions/setup-node@v4
-        with:
-          node-version: "22"
-          cache: "npm"
-
-      - name: Install dependencies
-        run: npm ci
-
-      - name: Run type checking
-        run: npm run typecheck
-
-  lint:
-    name: ESLint
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@v4
-
-      - name: Setup Node.js
-        uses: actions/setup-node@v4
-        with:
-          node-version: "22"
-          cache: "npm"
-
-      - name: Install dependencies
-        run: npm ci
-
-      - name: Run linting
-        run: npm run lint
-
-  format:
-    name: Prettier Format Check
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@v4
-
-      - name: Setup Node.js
-        uses: actions/setup-node@v4
-        with:
-          node-version: "22"
-          cache: "npm"
-
-      - name: Install dependencies
-        run: npm ci
-
-      - name: Check formatting
-        run: npm run format
-
-  build:
-    name: Build
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@v4
-
-      - name: Setup Node.js
-        uses: actions/setup-node@v4
-        with:
-          node-version: "22"
-          cache: "npm"
-
-      - name: Install dependencies
-        run: npm ci
-
-      - name: Build application
-        run: |
-          # Load CI environment variables for build-time evaluation
-          set -a
-          source .env.ci
-          set +a
-          npm run build
-
-  test:
-    name: Unit & Integration Tests
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@v4
-
-      - name: Setup Node.js
-        uses: actions/setup-node@v4
-        with:
-          node-version: "22"
-          cache: "npm"
-
-      - name: Install dependencies
-        run: npm ci
-
-      - name: Run tests
-        run: npm test
-
-  coverage:
-    name: Test Coverage
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@v4
-
-      - name: Setup Node.js
-        uses: actions/setup-node@v4
-        with:
-          node-version: "22"
-          cache: "npm"
-
-      - name: Install dependencies
-        run: npm ci
-
-      - name: Run tests with coverage
-        run: npm run test:coverage
-
-      - name: Upload coverage reports
-        uses: actions/upload-artifact@v4
-        if: always()
-        with:
-          name: coverage-report
-          path: coverage/
-          retention-days: 30
-
-  e2e:
-    name: E2E Smoke Tests
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@v4
-
-      - name: Setup Node.js
-        uses: actions/setup-node@v4
-        with:
-          node-version: "22"
-          cache: "npm"
-
-      - name: Install dependencies
-        run: npm ci
-
-      - name: Install Playwright browsers
-        run: pnpm exec playwright install --with-deps chromium
-
-      - name: Run smoke tests
-        run: npm run smoke
-
-      - name: Upload Playwright report
-        uses: actions/upload-artifact@v4
-        if: failure()
-        with:
-          name: playwright-report
-          path: playwright-report/
-          retention-days: 30
-
-      - name: Upload Playwright trace
-        uses: actions/upload-artifact@v4
-        if: failure()
-        with:
-          name: playwright-trace
-          path: test-results/
-          retention-days: 30
-
-  gitleaks:
-    name: Gitleaks Secret Scan
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@v4
-        with:
-          fetch-depth: 0
-
-      - name: Run Gitleaks
-        uses: gitleaks/gitleaks-action@v2
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-
-  npm-audit:
-    name: npm audit
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@v4
-
-      - name: Setup Node.js
-        uses: actions/setup-node@v4
-        with:
-          node-version: "22"
-          cache: "npm"
-
-      - name: Run npm audit
-        run: npm audit --audit-level=high
diff --git a/.gitignore b/.gitignore
index 4ae585b88..306864800 100644
--- a/.gitignore
+++ b/.gitignore
@@ -60,6 +60,7 @@ logs/
 
 # Temporary files
 /tmp
+.gemini/tmp/
 .recycle_bin/
 /data/
 #index.html
diff --git a/docs/plans/update_creating_skills.md b/docs/plans/update_creating_skills.md
new file mode 100644
index 000000000..3af45a49a
--- /dev/null
+++ b/docs/plans/update_creating_skills.md
@@ -0,0 +1,47 @@
+# Plan for updating creating-skills/SKILL.md
+
+## Goal
+
+Update the "Writing Skills" skill to be compatible with Gemini CLI's new Agent Skills support and the agentskills.io standard.
+
+## References
+
+- Gemini CLI PR #15698: Added tiered skill discovery (`~/.gemini/skills`, `.gemini/skills`) and `SkillManager`.
+- Agent Skills Standard: Uses `SKILL.md` with YAML frontmatter.
+
+## Proposed Changes
+
+1.  **Frontmatter**:
+    - Keep `name` and `description`.
+    - Keep `version` and `languages`.
+    - Keep `when_to_use` as it helps the agent understand context, even if strictly `description` is the primary discovery key in the CLI currently.
+
+2.  **File Paths & Locations**:
+    - Change `${SUPERPOWERS_SKILLS_ROOT}` to `.gemini/skills` (project-local) and `~/.gemini/skills` (global).
+    - Explain the tiered discovery: Project overrides User.
+
+3.  **Terminology**:
+    - "Claude" -> "Gemini" / "Agent".
+    - "Superpowers" -> "Gemini CLI".
+    - `find-skills` -> `/skills` command or automatic discovery.
+
+4.  **Structure Definition**:
+    - Update the `SKILL.md` template to match the standard.
+    - Emphasize that `SKILL.md` is the entry point.
+
+5.  **Workflow**:
+    - Update the "Discovery Workflow" to reflect how Gemini CLI loads skills (automatic discovery based on placement).
+
+## Step-by-Step Implementation
+
+1.  **Read** the file `.gemini/skills/creating-skills/SKILL.md`.
+2.  **Replace** the Overview section to explain Gemini CLI skill locations.
+3.  **Update** the Directory Structure section to show `.gemini/skills/`.
+4.  **Update** "Discovery Workflow" to remove `find-skills` tool reference and explain the new discovery mechanism.
+5.  **Refine** the "Claude Search Optimization" section to "Agent Search Optimization" (ASO) or similar, focusing on how Gemini will read the `SKILL.md`.
+6.  **Verify** the TDD process remains applicable (it does, as it's about content quality).
+
+## Verification
+
+- The file should describe how to create a skill that _works_ in Gemini CLI.
+- The pathing instructions must be correct for the user's OS (Linux) and Gemini CLI structure.
diff --git a/src/components/ui/button.tsx b/src/components/ui/button.tsx
index 86ec00acb..24e88d1e7 100644
--- a/src/components/ui/button.tsx
+++ b/src/components/ui/button.tsx
@@ -67,8 +67,7 @@ const Button = React.forwardRef<HTMLButtonElement, ButtonProps>(
         className={cn(buttonVariants({ variant, size, className }))}
         ref={ref}
         {...props}
-        // eslint-disable-next-line @typescript-eslint/prefer-nullish-coalescing -- We need logical OR here to correctly handle disabled={false} || loading={true}
-        disabled={props.disabled || loading}
+        disabled={!!props.disabled || loading}
       >
         {loading && <Loader2 className="mr-2 size-4 animate-spin" />}
         {props.children}