diff --git a/.claude/skills/README.md b/.claude/skills/README.md
new file mode 100644
index 000000000000..9b21f704989e
--- /dev/null
+++ b/.claude/skills/README.md
@@ -0,0 +1,36 @@
+# Claude Skills Directory
+
+This directory contains Claude Code skills for OpenSearch Dashboards development. 
+
+**Multi-Platform Support**: Skills are available for both Claude Code and Kiro users through corresponding prompt formats.
+
+## Available Skills
+
+### `resolve-cve` ([resolve_cve.md](resolve_cve.md))
+Automatically identify and resolve security vulnerabilities (CVEs) in project dependencies.
+
+- **Documentation**: [resolve_cve_readme.md](resolve_cve_readme.md)
+- **Claude Usage**: `/resolve-cve [--cve_id CVE-2024-12345]`
+- **Kiro Usage**: Also available as Kiro prompt at [.kiro/prompts/resolve_cve.md](../../.kiro/prompts/resolve_cve.md)
+- **Purpose**: Scans GitHub issues for CVEs, verifies presence in codebase, attempts automated fixes
+
+## Adding New Skills
+
+When adding new skills to this directory:
+
+1. **Claude skill file**: `skill-name.md` (the actual skill definition)
+2. **Kiro prompt file**: `../../.kiro/prompts/skill-name.md` (Kiro-format prompt that delegates to Claude skill)
+3. **Documentation**: `skill_name_readme.md` (comprehensive user guide)  
+4. **Update this README**: Add entry to "Available Skills" section with both usage formats
+
+## Skill Development Guidelines
+
+- Follow the existing skill template format with frontmatter
+- Include clear usage examples and error handling
+- Document all parameters and expected outputs
+- Test thoroughly before committing
+- Keep skills focused on specific, well-defined tasks
+
+---
+
+For general Claude Code documentation, see [CLAUDE.md](../../CLAUDE.md) in the project root.
\ No newline at end of file
diff --git a/.claude/skills/resolve_cve.md b/.claude/skills/resolve_cve.md
new file mode 100644
index 000000000000..c695d14cc1ba
--- /dev/null
+++ b/.claude/skills/resolve_cve.md
@@ -0,0 +1,207 @@
+---
+name: resolve-cve
+description: Automatically resolve CVEs by checking GitHub issues, verifying presence in codebase, and attempting various remediation strategies while documenting the process.
+arguments:
+  - name: cve_id
+    description: Specific CVE identifier to resolve (e.g., CVE-2024-12345). If not provided, will search for all open CVE issues.
+    required: false
+---
+
+# CVE Resolution Skill
+
+Automatically resolve CVEs by checking GitHub issues, verifying presence in codebase, and attempting various remediation strategies while documenting the process.
+
+## Usage
+
+```
+/resolve-cve [--cve_id CVE-2024-12345]
+```
+
+If no CVE ID is provided, will search for all open CVE issues.
+
+## Process
+
+You are a CVE resolution specialist for OpenSearch Dashboards. Your goal is to automatically identify and resolve security vulnerabilities in the project's dependencies.
+
+**Important**: 
+- Always create reports in `tmp/` directory (gitignored) to avoid committing temporary files
+- Clean existing tmp files at start of each run to prevent stale data in PR descriptions
+
+### Step 1: Identify CVEs
+
+1. **Search GitHub Issues**: Look for open CVE issues in https://github.com/opensearch-project/OpenSearch-Dashboards/issues
+
+   - Use GitHub API or web search to find issues with "CVE" in title
+   - Filter for open issues only
+   - Extract CVE numbers and affected package information
+
+2. **Parse CVE Details**: For each CVE found, extract:
+   - CVE identifier (e.g., CVE-2024-12345)
+   - Affected package name and vulnerable version range
+   - Recommended safe version
+   - Severity level
+
+### Step 2: Verify Presence in Codebase
+
+1. **Check package.json**: Search for the vulnerable package in package.json files
+2. **Check yarn.lock**: Verify if vulnerable versions are actually installed
+3. **Analyze dependency tree**: Use `yarn why <package>` to understand why the package is included
+
+### Step 3: Remediation Strategies
+
+Try the following strategies in order until successful:
+
+#### Strategy A: Direct Package Update (if in package.json)
+
+1. If vulnerable package is explicitly declared in package.json:
+   - Update version to safe version with caret (^)
+   - Run `yarn osd bootstrap` to update lock file and verify build
+
+#### Strategy B: Lock File Manipulation (if transitive dependency)
+
+1. If package is NOT in package.json (transitive dependency):
+   - Delete vulnerable package entries from yarn.lock
+   - Run `yarn osd bootstrap` to regenerate with latest versions
+   - Check if vulnerability is resolved
+
+#### Strategy C: Peer Dependency Resolution
+
+1. If Strategy B fails due to peer dependency constraints:
+   - Identify which dependency requires the vulnerable version
+   - Delete that dependency's yarn.lock entries too
+   - Run `yarn osd bootstrap` again
+   - May need to update parent dependency version
+
+#### Strategy D: Resolutions (last resort)
+
+1. If other strategies fail:
+   - Add yarn resolutions to package.json to force safe version
+   - Use the least invasive resolution path possible
+   - Example: If package "a" has a CVE and the vulnerable version comes from dependency "b", use a targeted resolution like `"**/b/a": "^3.5.3"` instead of a global `"a": "^3.5.3"`
+   - Document why the resolution was needed and which dependency path required it
+
+### Step 4: Verification
+
+After each remediation attempt:
+
+1. Run `yarn osd bootstrap` to ensure build succeeds
+2. Run `yarn audit` to check if vulnerability is resolved
+3. Run basic smoke tests if available
+4. Verify no new vulnerabilities were introduced
+
+### Step 5: Documentation
+
+Create a PR-ready description in `tmp/cve-pr-description.md` by **dynamically using the current GitHub PR template**:
+
+1. **Read the current PR template**: Load `.github/pull_request_template.md` to get the latest format
+2. **Parse template structure**: Identify sections like Description, Issues Resolved, Testing, Changelog, etc.
+3. **Fill in CVE-specific content**: Replace template placeholders with actual CVE resolution details
+4. **Auto-extract GitHub Issue Numbers**: Parse CVE issue URLs and add `closes #[number]` entries
+
+**Content Mapping Strategy**:
+
+- **Description section**: Fill with CVE summary, affected packages, resolution strategy
+- **Issues Resolved section**: Auto-populate with `closes #[issue-number]` from found CVE issues
+- **Testing section**: Add CVE-specific verification commands
+- **Changelog section**: Generate appropriate changelog entry for security fixes
+- **Checklist section**: Mark relevant items as completed based on resolution results
+
+**Dynamic Template Approach**:
+
+```bash
+# Always use current template as base
+cp .github/pull_request_template.md tmp/cve-pr-description.md
+
+# Then programmatically fill in sections with CVE data:
+# - Replace "<!-- Describe what this change achieves-->" with CVE details
+# - Replace "<!-- List any issues this PR will resolve -->" with closes #123
+# - Fill testing section with CVE verification steps
+# - Add appropriate changelog entry
+```
+
+This ensures the skill **always respects the current PR template format**, even if it changes in the future - no manual skill updates required!
+
+### Step 6: Setup and Failure Documentation
+
+**Setup tmp/ directory and generate PR description**:
+
+```bash
+mkdir -p tmp
+echo "tmp/" >> .gitignore  # if not already present
+
+# IMPORTANT: Clean any existing tmp files to avoid stale data
+rm -f tmp/cve-pr-description.md tmp/cve-failure-report.md
+
+# Use current GitHub PR template as base (future-proof!)
+cp .github/pull_request_template.md tmp/cve-pr-description.md
+```
+
+**Why we clean tmp files first**:
+- **Prevents stale data**: Ensures fresh reports with current CVE status
+- **Accurate issue refs**: Avoids outdated "closes #123" references  
+- **No confusion**: Eliminates mixing data from multiple CVE runs
+- **Current state**: Always reflects the actual dependency versions being fixed
+
+**Template Content Replacement Guide**:
+
+1. **Description section**: Replace `<!-- Describe what this change achieves-->` with:
+
+   ```
+   Resolves security vulnerabilities in project dependencies:
+   - **CVE-[ID]** ([Severity]): [Brief description]
+   - **Package**: [affected-package@version] → [safe-version]
+   - **Resolution Strategy**: [Strategy used]
+   ```
+
+2. **Issues Resolved section**: Replace `<!-- List any issues... -->` with:
+
+   ```
+   - closes #[issue-number]  (auto-extracted from GitHub CVE issues)
+   ```
+
+3. **Testing section**: Replace `<!-- Please provide detailed steps... -->` with CVE-specific validation steps
+
+4. **Changelog section**: Add appropriate security fix entry following existing format
+
+This approach ensures the skill **automatically adapts** to any future PR template changes without requiring skill updates.
+
+If CVE cannot be automatically resolved, document in `tmp/cve-failure-report.md`:
+
+- Which strategies were attempted
+- Specific error messages encountered
+- Dependency conflicts preventing resolution
+- Recommended manual intervention steps
+- Upstream dependency update requirements
+- GitHub issue numbers for tracking
+
+## Error Handling
+
+- Always backup original package.json and yarn.lock before making changes
+- If any step fails, restore backups and document the failure
+- Never leave the project in a broken state
+- Provide clear next steps for manual resolution
+
+## Constraints
+
+- Only modify dependency versions, never remove required dependencies
+- Maintain compatibility with Node.js version requirements
+- Preserve existing functionality
+- Follow semver best practices for version updates
+- Document all changes for PR review
+
+## Success Criteria
+
+A CVE is considered resolved when:
+
+1. Vulnerable package version is no longer present
+2. Project builds successfully (`yarn osd bootstrap` passes)
+3. All tests pass (`yarn test:jest` passes)
+4. No new vulnerabilities introduced (`yarn audit` clean for resolved CVE)
+5. PR-ready documentation created in `tmp/cve-pr-description.md`
+6. GitHub issue numbers auto-extracted and added as `closes #123` references
+
+## Output Files
+
+- `tmp/cve-pr-description.md` - PR description ready to copy/paste, **dynamically generated from current `.github/pull_request_template.md`**
+- `tmp/cve-failure-report.md` - Failure analysis (if resolution fails), using same template approach
+- Both files automatically adapt to future PR template changes - no skill maintenance required!
diff --git a/.claude/skills/resolve_cve_readme.md b/.claude/skills/resolve_cve_readme.md
new file mode 100644
index 000000000000..08da1d00581f
--- /dev/null
+++ b/.claude/skills/resolve_cve_readme.md
@@ -0,0 +1,230 @@
+# resolve-cve Skill Documentation
+
+Comprehensive guide for the **`resolve-cve`** skill (`resolve_cve.md`).
+
+Automated tool for identifying and resolving security vulnerabilities (CVEs) in OpenSearch Dashboards dependencies.
+
+## 🔧 What it does
+
+The CVE resolution skill automatically:
+- 🔍 **Finds open CVE issues** from GitHub  
+- 🔎 **Verifies which affect your code** by analyzing package.json and yarn.lock
+- 🛠️ **Attempts multiple fix strategies** (direct updates, lock manipulation, resolutions)
+- ✅ **Validates the fixes work** by running builds and audits
+- 📋 **Documents everything** in a comprehensive report for your PR
+
+## 🚀 Usage
+
+### Basic Commands
+
+```bash
+# Scan and fix ALL open CVEs
+/resolve-cve
+
+# Target a specific CVE
+/resolve-cve --cve_id CVE-2025-54798
+
+# Get help on available options
+/resolve-cve --help
+```
+
+### Example Workflow
+
+1. **Run the skill:**
+   ```bash
+   /resolve-cve --cve_id CVE-2025-54798
+   ```
+
+2. **Review the generated `tmp/cve-pr-description.md`** for PR-ready content
+
+3. **Test your changes:**
+   ```bash
+   yarn osd bootstrap
+   yarn test:jest  # optional
+   ```
+
+4. **Copy `tmp/cve-pr-description.md` content to your PR** - it's formatted and ready!
+
+## 📂 Output Files
+
+After running the skill, look in the `tmp/` directory for:
+
+- **`tmp/cve-pr-description.md`** 📋 - **Main output**: Copy this content into your PR description
+- **`tmp/cve-failure-report.md`** ⚠️ - **If resolution fails**: Manual steps and error analysis
+
+> **💡 Tip**: The PR description follows GitHub's template format and includes auto-generated "closes #123" references!
+
+## 📋 Prerequisites
+
+Make sure you have:
+- ✅ **Clean working directory** (commit any pending changes)
+- ✅ **Yarn installed** (the skill uses `yarn osd bootstrap`)
+- ✅ **GitHub access** (for searching CVE issues)
+- ✅ **Node.js 20+** (required by OpenSearch Dashboards)
+
+## 📂 File Management
+
+**Important**: The skill automatically cleans `tmp/` files at the start of each run to prevent stale data:
+
+```bash
+# These files are removed before generating new reports
+rm -f tmp/cve-pr-description.md tmp/cve-failure-report.md
+```
+
+**Why this matters**:
+- 🚫 **Prevents outdated PR content** - No risk of copying old CVE info
+- ✅ **Ensures current data** - Reports always reflect actual dependency state  
+- 🎯 **Accurate issue references** - Correct "closes #123" GitHub issue numbers
+- 🔄 **Clean slate each run** - No confusion from previous CVE resolution attempts
+
+If you need to preserve previous reports, copy them elsewhere before running the skill again.
+
+## 🔍 How it works
+
+### Step 1: CVE Discovery
+- Searches https://github.com/opensearch-project/OpenSearch-Dashboards/issues
+- Filters for open issues containing "CVE"
+- Extracts vulnerability details (package, versions, severity)
+
+### Step 2: Impact Analysis  
+- Checks if vulnerable packages exist in your `package.json`
+- Analyzes `yarn.lock` for vulnerable transitive dependencies
+- Uses `yarn why <package>` to understand dependency chains
+
+### Step 3: Smart Remediation
+The skill tries multiple strategies in order:
+
+**Strategy A: Direct Updates** 
+- For packages explicitly in `package.json`
+- Updates to safe versions with caret ranges
+
+**Strategy B: Lock File Manipulation**
+- For transitive dependencies  
+- Removes vulnerable entries, lets yarn regenerate with latest
+
+**Strategy C: Parent Dependency Updates**
+- When Strategy B fails due to constraints
+- Updates parent dependencies that pull in vulnerable packages  
+
+**Strategy D: Yarn Resolutions** ⭐
+- Last resort for stubborn vulnerabilities
+- Forces specific safe versions using targeted resolutions
+- Example: `"**/selenium-webdriver/tmp": "^0.2.4"`
+
+### Step 4: Validation
+- ✅ Runs `yarn osd bootstrap` to ensure build works
+- ✅ Checks `yarn audit` to confirm CVE is resolved  
+- ✅ Verifies no new vulnerabilities introduced
+- ✅ Creates backup files for easy rollback
+
+## 📄 Generated Documentation
+
+The skill creates comprehensive documentation in the `tmp/` directory:
+
+**Primary Output: `tmp/cve-pr-description.md`** - Ready-to-copy PR description following GitHub template format
+
+**Content includes:**
+
+```markdown
+### Description
+
+Resolves security vulnerability in project dependencies:
+- **CVE-2025-54798** (Low): Vulnerability in tmp package
+- **Package**: tmp@0.0.30 → tmp@0.2.5
+
+### Issues Resolved
+- closes #10376
+
+## Testing the changes
+[CVE-specific verification commands]
+
+## Changelog
+- fix: Resolve CVE-2025-54798 in tmp dependency
+```
+
+**Failure Reports: `tmp/cve-failure-report.md`** - Detailed analysis if resolution fails
+
+## 🛡️ Safety Features
+
+- **🔙 Automatic backups**: Creates `*.bak` files before any changes
+- **🚫 Never breaks builds**: Validates changes before proceeding  
+- **📝 Complete audit trail**: Documents every action taken
+- **🔄 Easy rollback**: Restore from backups if needed
+- **🎯 Targeted fixes**: Uses least invasive resolution strategies
+
+## 🐛 Troubleshooting
+
+### "Unknown skill: resolve-cve"
+The skill needs to be in your global skills directory:
+```bash
+# Check if skill exists
+ls ~/.claude/skills/resolve-cve/
+
+# If not, skills may be project-local only
+ls .claude/skills/resolve-cve.md
+```
+
+### "Build fails after CVE fix"
+1. **Restore backups:**
+   ```bash
+   mv package.json.bak package.json
+   mv yarn.lock.bak yarn.lock
+   yarn osd bootstrap
+   ```
+
+2. **Check `tmp/cve-failure-report.md`** for manual steps required
+
+3. **Try a different strategy** - some CVEs need manual intervention
+
+### "CVE still shows in audit"
+- Check if multiple packages have the same vulnerability
+- Some audit tools cache results - try `yarn audit --force`
+- Verify you're checking the right CVE ID
+
+### "Skill doesn't find CVE issues"  
+- Ensure you have internet access for GitHub API calls
+- Check if the CVE issue still exists and is open
+- Try searching GitHub manually: "repo:opensearch-project/OpenSearch-Dashboards CVE"
+
+## 💡 Pro Tips
+
+1. **Run regularly**: Use `/resolve-cve` weekly to catch new vulnerabilities early
+
+2. **Test thoroughly**: Always run the full test suite after CVE fixes:
+   ```bash
+   yarn test:jest
+   yarn test:jest_integration  
+   ```
+
+3. **Review yarn resolutions**: If Strategy D was used, periodically check if resolutions can be removed after dependency updates
+
+4. **Batch related CVEs**: If multiple CVEs affect the same package, fix them together
+
+5. **Keep backups**: The `*.bak` files are your safety net - don't delete them until your PR is merged
+
+## 🔗 Related Commands
+
+```bash
+# Check current vulnerabilities
+yarn audit --level moderate
+
+# Find why a package is installed  
+yarn why <package-name>
+
+# Update all dependencies (careful!)
+yarn upgrade-interactive --latest
+
+# Clean build after fixes
+yarn osd clean && yarn osd bootstrap
+```
+
+## 📞 Getting Help
+
+- **Skill issues**: Check the skill file at `.claude/skills/resolve-cve.md`
+- **CVE questions**: Review files in `tmp/` directory (cve-pr-description.md, cve-failure-report.md)
+- **Build problems**: See OpenSearch Dashboards troubleshooting in `CLAUDE.md`
+- **Security concerns**: Consult with security team before merging CVE fixes
+
+---
+
+**Happy CVE hunting! 🛡️**
\ No newline at end of file
diff --git a/.gitignore b/.gitignore
index 8cf39608517c..473d4c70174d 100644
--- a/.gitignore
+++ b/.gitignore
@@ -4,7 +4,13 @@
 /.opensearch
 /.chromium
 /package.json.bak
+*.bak
 .claude/
+!.claude/skills/
+
+# Temporary files and directories
+tmp/
+.claude/settings.local.json
 CLAUDE.local.md
 .DS_Store
 .node_binaries
@@ -50,7 +56,8 @@ selenium
 package-lock.json
 .yo-rc.json
 .vscode
-.kiro
+.kiro/
+!.kiro/prompts/
 *.sublime-*
 npm-debug.log*
 .tern-project
diff --git a/.kiro/prompts/resolve_cve.md b/.kiro/prompts/resolve_cve.md
new file mode 100644
index 000000000000..5dba00c924ea
--- /dev/null
+++ b/.kiro/prompts/resolve_cve.md
@@ -0,0 +1,64 @@
+---
+name: resolve-cve
+description: Automatically resolve CVEs by checking GitHub issues, verifying presence in codebase, and attempting various remediation strategies while documenting the process.
+parameters:
+  - name: cve_id
+    description: Specific CVE identifier to resolve (e.g., CVE-2024-12345). If not provided, will search for all open CVE issues.
+    required: false
+    type: string
+---
+
+# CVE Resolution (Kiro)
+
+You are a CVE resolution specialist for OpenSearch Dashboards. Your task is to automatically identify and resolve security vulnerabilities in the project's dependencies.
+
+## Instructions
+
+**IMPORTANT**: Read and follow the comprehensive CVE resolution process defined in `.claude/skills/resolve_cve.md`. That file contains the complete, detailed instructions for:
+
+1. **CVE Discovery**: How to search GitHub issues and parse CVE details
+2. **Impact Analysis**: Verifying presence in codebase via package.json and yarn.lock
+3. **Remediation Strategies**: Four different approaches (A-D) to try in order
+4. **Verification**: Build validation and audit confirmation
+5. **Documentation**: Creating PR-ready descriptions in `tmp/` directory
+
+## Key Requirements
+
+- **Follow the exact process** outlined in `.claude/skills/resolve_cve.md`
+- **Use the `tmp/` directory** for all generated reports to avoid git conflicts
+- **Clean existing tmp files** at start of each run to prevent stale data
+- **Generate PR descriptions** using the dynamic GitHub template approach
+- **Auto-extract issue numbers** and add appropriate "closes #123" references
+
+## Parameters
+
+- `cve_id` (optional): Specific CVE to target (e.g., CVE-2024-12345)
+- If no CVE ID provided, process all open CVE issues
+
+## Output
+
+Generate these files in `tmp/` directory:
+- `tmp/cve-pr-description.md` - Ready-to-copy PR description
+- `tmp/cve-failure-report.md` - Analysis if resolution fails
+
+## Process Overview
+
+Refer to `.claude/skills/resolve_cve.md` for the complete step-by-step process, including:
+
+- **Strategy A**: Direct package updates (if in package.json)
+- **Strategy B**: Lock file manipulation (transitive dependencies) 
+- **Strategy C**: Parent dependency updates (when constraints prevent B)
+- **Strategy D**: Yarn resolutions (last resort, most reliable)
+
+## Success Criteria
+
+A CVE is resolved when:
+1. Vulnerable package version no longer present
+2. Project builds successfully (`yarn osd bootstrap`)
+3. Tests pass (`yarn test:jest`) 
+4. Audit clean for the target CVE
+5. PR-ready documentation generated in `tmp/`
+
+---
+
+**Execute the CVE resolution process exactly as specified in `.claude/skills/resolve_cve.md`**
\ No newline at end of file
diff --git a/changelogs/fragments/11650.yml b/changelogs/fragments/11650.yml
new file mode 100644
index 000000000000..f1ba9e4d9f46
--- /dev/null
+++ b/changelogs/fragments/11650.yml
@@ -0,0 +1,2 @@
+feat:
+- Add automated CVE resolution skill with multi-platform support ([#11650](https://github.com/opensearch-project/OpenSearch-Dashboards/pull/11650))
\ No newline at end of file