Skill UI

A governance skill designed to secure AI agent tool calls using advanced authorization policies. It implements Cedar policies to control which tools agents can access and under what conditions. Key features include running in 'Shadow mode' for safe observation, 'Enforce mode' for strict blocking, and generating cryptographically signed Ed25519 receipts to provide an immutable, tamper-evident audit trail for every policy decision.

A comprehensive toolset for debugging and testing web applications in live browsers using Chrome DevTools MCP. It allows agents to inspect the DOM, capture console errors, analyze network traffic, profile performance, and verify visual output, bridging the gap between static analysis and real runtime execution. Essential for UI debugging and end-to-end testing.

Automates CI/CD pipeline setup, from linting and type checks to integration, E2E, security, and deployment stages, plus feedback loops for AI agents and preview deployments.

Guides AI agents through analyzing working code, spotting overly complex structures, then refactoring for clarity while preserving exact behavior and project conventions, especially after reviews or high-pressure edits.

Guides writing documentation and ADRs to capture why architectural or API choices were made, when to comment, and how to structure README/API info so future engineers and agents understand the rationale and trade-offs.

This guide details best practices for using `lark-cli` to interact with Feishu resources. It covers initial configuration, switching between user and bot identities, and crucial authentication flows for AI agents. It provides detailed instructions on handling permission errors, implementing secure split-flow authentication, and managing high-risk write operations to ensure data integrity and user consent.

Manage Lark tasks, tasklists, and task agents. Supports creating/updating tasks, assigning members, adding subtasks, uploading attachments, and tracking progress.

Provides OpenClaw agents with local persistent memory via MemOS, managing installation, configuration, upgrades, and a memory viewer for secure offline workflows.

Guides agents through researching Microsoft technologies via Learn MCP docs, then scaffolds hybrid skills that store essential knowledge locally while enabling dynamic lookups for depth.

This skill allows agents to ingest external resources (local files, URLs, repositories) using `ov add-resource`, and to create long-lasting searchable memories from conversations using `ov add-memory`. It is essential for building a comprehensive, persistent knowledge base for the OpenViking context database, ensuring that learned insights and external data remain available across sessions and interactions. Use when external knowledge injection or context persistence is required.

Use the OpenViking CLI to search memories, skills, and resources with semantic and literal queries across directories, surfacing context-aware results whenever agents need stored knowledge or documentation.

This mode simulates a P9 Tech Lead role, specializing in strategic project management and complex task orchestration. It does not write code itself, but focuses on generating detailed, structured Task Prompts (six elements) and managing the entire workflow delivery from multiple specialized agents (P8 teams). Ideal for coordinating large, complex projects requiring strategic decomposition and coordination of three or more parallel agents.