kg

Read the Code!

Perform a thorough and detailed code review

Use sub-agents to augment task files in parallel

Use sub-agents to execute task files in parallel

Use sub-agents to qa code review task files in parallel

Deep codebase researcher with KODEGEN tools. WHEN User asks "investigate", "analyze codebase", "find all X", "research how Y works", needs multi-file investigation. WHEN NOT Simple file reads, single searches, known file locations.

Autonomous browser agent for complex multi-step web tasks. WHEN: User needs complex web automation, multi-page workflows, form filling sequences, tasks requiring multiple browser actions. WHEN NOT: Simple single-action tasks like one click (use browser_click), one navigation (use browser_navigate), one extraction (use browser_extract_text).

Spawn and manage Claude sub-agents for parallel or delegated tasks. WHEN: User needs parallel operations, multi-step autonomous research, or wants to delegate complex tasks to sub-agents. Use when task can be parallelized or requires independent exploration. WHEN NOT: Simple single-step operations, sequential dependencies, specific file reads (use fs_read_file), writing code directly (do it yourself).

Memory system with three MCP tools: mcp__plugin_kg_kodegen__memory_list_libraries (discover namespaces), mcp__plugin_kg_kodegen__memory_memorize (store with embeddings), mcp__plugin_kg_kodegen__memory_recall (semantic search). Use for storing/retrieving knowledge by meaning. WHEN: User says "remember this", "recall", "what did I save", "my notes", "find my knowledge about". WHEN NOT: File operations (use fs_*), exact keyword search (use fs_search).

Advanced reasoning with search strategies (beam search, MCTS). WHEN: Complex problem-solving requiring exploration of multiple solution paths, optimization problems, decision trees, when you need scored/ranked reasoning paths. WHEN NOT: Simple linear reasoning (use sequential_thinking), trivial problems, when branching isn't needed.

Web crawling with Tantivy full-text search indexing. Supports crawl, search, and auto-crawl. WHEN: User wants to "scrape a website", "crawl documentation", "search crawled content", "index a site". WHEN NOT: Single page fetch (use browser_navigate), web search (use web_search).

2 hooks across 2 events