keysersoose/claude-agent-builder
Skill Claude CodeCodex
Build custom AI agents in Claude Code from a user's problem statement. This skill analyzes the user's use case, asks smart clarifying questions, researches the internet for similar agents (GitHub repos, blogs, Claude Code community patterns), and then architects and builds production-ready Claude Code agents …
Agent
Task planner for complex changes. Use PROACTIVELY when a task touches 3+ files, involves a new feature (not a bug fix), or requires understanding how components interact. Invoke BEFORE writing code. If the task is a simple single-file change, skip this agent entirely.
Agent
Code reviewer. Use before any git commit, when validating implementations, or when asked to review a PR or diff. Focuses on bugs and security, not style.
Command
Session boot sequence. Run at the start of every session to load memory and verify rules.
Command
Review the learning system and promote/prune rules. Run weekly or when corrections seem repetitive.
Command
Pre-commit review pipeline. Runs checks then reviews the diff for bugs, security, and performance issues.
Skill Claude CodeCodex
Source discovery - find high-quality candidate URLs with Google operators rendered in a real Chrome session, multi-query strategies, and source-quality triage, emitting deduplicated lead JSON. Use when a task needs to locate official docs, papers, code repositories, community evidence, recent posts, or competitive…
Skill Claude CodeCodex
Crawl4AI capture - turn web pages, including JavaScript-rendered ones, into LLM-ready markdown or structured CSS/JSON extractions, with batch crawling, screenshots/PDFs, and extraction-quality checks. Use when a task needs page content beyond what a simple fetch or browser read returns, or when deep-research captures…
Skill Claude CodeCodex
Deep research - end-to-end source-backed workflow that turns a research question into an agent-authored, cited brief with raw source captures. Use when a task needs a literature review, technical investigation, market or competitive intelligence, current-state analysis, any answer whose claims must be traceable to…
Skill Claude CodeCodex
Ingest a source into the global LLM wiki (the knowledge base path defined in your global CLAUDE.md) — Karpathy's first LLM-wiki operation. Takes a URL, file, export, or pasted material; produces a cited wiki page, updates the index and related pages, and logs the ingest. Use when asked to "ingest this", "add to the…
Skill Claude CodeCodex
Lint the global LLM wiki (the knowledge base path defined in your global CLAUDE.md) — the maintenance half of the Karpathy LLM-wiki loop. Checks contradictions, stale claims, orphan pages, index freshness, and broken links, then fixes what it can and reports what needs the user. Use when asked to "lint the wiki"…