ari-anthropic-monitor

name	ari-anthropic-monitor
description	Monitor official Anthropic releases for ARI improvements - VERIFIED SOURCES ONLY
triggers	["check anthropic updates","new claude features","anthropic releases","what's new from anthropic"]

ARI Anthropic Monitor

Purpose

Track official Anthropic releases to keep ARI current with the latest capabilities, security patches, and best practices. Only uses verified Anthropic sources.

Verified Sources (ONLY these are trusted)

Official Anthropic Sources

Source	URL	Trust Level
Anthropic News	https://www.anthropic.com/news	SYSTEM
API Release Notes	https://docs.anthropic.com/en/release-notes	SYSTEM
Claude Code Releases	https://github.com/anthropics/claude-code/releases	SYSTEM
Anthropic SDK	https://github.com/anthropics/anthropic-sdk-python	SYSTEM
Claude Blog	https://claude.com/blog	SYSTEM
Anthropic Research	https://www.anthropic.com/research	SYSTEM

Official Plugin Sources

Source	Trust Level	Verification
claude-plugins-official	VERIFIED	Anthropic-managed
anthropics/claude-code	VERIFIED	Anthropic repo
trailofbits/skills	VERIFIED	Security firm

Security Protocol

What This Skill WILL Do

• Fetch release notes from official sources
• Summarize new features relevant to ARI
• Identify security patches that apply
• Recommend verified plugin updates

What This Skill WILL NOT Do

• Install anything automatically
• Download from unverified sources
• Execute code from external sources
• Bypass ARI's security review

Monitoring Workflow

1. User invokes "check anthropic updates"

2. Fetch from VERIFIED sources only:
   - WebFetch → anthropic.com/news
   - WebFetch → docs.anthropic.com/release-notes
   - gh api → anthropics/claude-code releases

3. Parse and filter for ARI-relevant updates:
   - Claude API changes
   - Claude Code plugin system
   - Security advisories
   - Performance improvements
   - New capabilities

4. Present findings with security assessment:
   - Source verification status
   - Compatibility with ARI architecture
   - Security implications
   - Implementation complexity

5. User decides whether to proceed

Output Format

## Anthropic Updates Report
**Generated**: [timestamp]
**Sources Checked**: [list of verified sources]

### Security Advisories (PRIORITY)
[Any security-related updates]

### Claude API Updates
[New API features or changes]

### Claude Code Updates
[Plugin system, hooks, commands]

### Research & Papers
[Relevant research that could improve ARI]

### Recommendations for ARI
| Update | Priority | Security Impact | Requires Review |
|--------|----------|-----------------|-----------------|
| ... | ... | ... | ... |

Integration with ARI Governance

All recommendations flow through:

1. Guardian Agent - Threat assessment
2. Arbiter - Constitutional compliance
3. Council - Approval for significant changes

Trust Verification

Before any action:

// Verify source is in trusted list
const TRUSTED_SOURCES = [
  'anthropic.com',
  'docs.anthropic.com',
  'claude.com',
  'github.com/anthropics',
  'github.com/trailofbits'
];

function verifySource(url: string): boolean {
  return TRUSTED_SOURCES.some(trusted => url.includes(trusted));
}

Example Interaction

User: "check anthropic updates"

Skill:
Checking verified Anthropic sources...

✅ anthropic.com/news - Connected
✅ docs.anthropic.com/release-notes - Connected
✅ github.com/anthropics/claude-code - Connected

## Anthropic Updates Report
Generated: 2026-01-28

### Security Advisories
None found.

### Claude API Updates
- Claude 3.5 Opus: Improved reasoning (Jan 15)
- Tool use enhancements (Jan 10)

### Claude Code Updates
- Plugin system v2.1: New hook types
- Marketplace security improvements

### Recommendations for ARI
| Update | Priority | Security | Action |
|--------|----------|----------|--------|
| Hook system v2.1 | Medium | Safe | Review |
| Tool use update | Low | Safe | Monitor |

No automatic installations. Review recommendations?