Las Mejores Habilidades de Claude del 23 de junio de 2026

HADS Documentation Standard

Standardizes technical documentation for optimal readability by both humans and AI models using the Human-AI Document Standard.

Protect MCP: Policy & Audit Enforcement

Secures Claude Code tool executions using Cedar policy enforcement and Ed25519 cryptographic signatures for verifiable audit trails.

Plugin Quality Methodology

Standardizes AI plugin and skill assessment through multi-layered scoring, statistical validation, and comprehensive quality rubrics.

Git Guard: No-Verify Blocker

Enforces repository quality gates by preventing AI agents from bypassing git pre-commit hooks and GPG signing requirements.

Signed Audit Trails for Claude Code

Implements cryptographically signed receipts and Cedar policy enforcement for AI agent tool calls to ensure tamper-evident auditability.

Fission Go Security Upgrader

Automates grouped Go dependency security updates and vulnerability fixes using a bisectable commit strategy.

Harness Loop

Automates long-running software development cycles by executing autonomous Plan-Work-Review loops with persistent state management.

Autonomous Development Loop

Orchestrates long-running, autonomous development cycles by managing persistent task execution and state across Claude Code sessions.

Harness Loop

Automates long-running development cycles by executing, reviewing, and coordinating task batches in a continuous background loop.

Miro Debug Bundle

Collects diagnostic evidence and API health data for troubleshooting Miro REST API v2 integrations.