Agent Skills Hub

guard

Epistemic safety analysis for JSON data in prompts to prevent LLM hallucinations and reasoning errors when handling incomplete or large-scale datasets.

log-focus-debug

Focus debug skill for DashPlayer: isolates log chains, injects temporary focus markers ([FOCUS:token]), and ensures clean removal of debug artifacts after task completion.

documentation

Write, structure, and maintain technical documentation like READMEs, API docs, runbooks, and architecture specs to keep your team aligned and informed.

memory-hygiene

Audit, prune, and maintain vector memory for Clawdbot. Prevents token waste, clears junk data, and automates memory hygiene via LanceDB maintenance.

doc-sync

Synchronizes and maintains CLAUDE.md and README.md documentation hierarchy across a repository to ensure consistent, just-in-time context for AI agents.

massive-context-mcp

Process massive files and large codebases (10M+ tokens) by recursively chunking, sub-querying, and aggregating results to overcome LLM context limits.

dlt-skill

Create, manage, and debug dlt (data load tool) pipelines for ingesting data from APIs, databases, and custom sources into destinations like DuckDB, BigQuery, and Snowflake.

Global Documentation

Ensures adherence to standardized global documentation patterns for technical projects, maintaining consistency across repositories and agent-based workflows.

ai-llm-patterns

Anthropic Claude integration patterns: streaming, RAG with pgvector, tool use, model selection (Haiku/Sonnet/Opus), prompt caching, and cost management for AI-powered engineering.

ReasoningBank Intelligence

Implement adaptive learning with ReasoningBank for pattern recognition, strategy optimization, and continuous improvement in AI agents.

doc-audit

Audit Packmind documentation by cross-referencing MDX files against the codebase to detect broken links, outdated CLI references, and missing coverage.

youtube-digest

Analyze YouTube videos with automated transcript extraction, AI-powered summarization, Korean translation, and interactive multi-level comprehension quizzes.