A personal collection of AI resources — skills, agents, MCP servers, etc. This is a living collection; more categories will be added as they're picked up.
install.sh symlinks every skill in skills/ and every agent in agents/ into ~/.claude/skills/ and ~/.claude/agents/, overriding anything already installed under the same name.
./install.shagents/dan-agent.md— routing target for/dan-mode; reads thedan-modeskill in full before any work.agents/comment-sicko.md— deranged comment-hater that hunts down and deletes narration, banners, and workaround comments.
skills/architect/— sketch types, signatures, and module structure before code, then stay in the loop while implementation fills in.skills/arena/— spawn N parallel candidates at the same task, pick a base, graft the strongest parts of the losers into it.skills/automate-me/— draft or revise a personal "-mode" skill capturing how the user works, via skill-creator + unslop.skills/blast-radius/— find what a change could break beyond the diff, and prove the one fact that makes it safe by running real code.skills/bro/— restate the last message in plain human language, no jargon.skills/coding-best-practice/— language-agnostic coding guidelines covering restraint and design, distilled from classic engineering literature plus common LLM failure modes.skills/create-verification-skill/— generate a project-local verification skill that drives an app the way a user does.skills/dan-mode/— Dan's agent style: concise responses, deliberate subagents, unslopped prose, simple code, verified work.skills/figure-it-out/— auditable playbook for large migrations or ambitious multi-part changes with no narrower playbook.skills/grilling/— grill the user relentlessly about a plan, decision, or idea to stress-test their thinking.skills/how/— explain subsystem architecture, runtime flow, and placement/ownership questions.skills/interrogate/— multiple LLM reviewers challenge a change from independent angles.skills/investment-buffett/— Warren Buffett's investment thinking system for stock/company analysis and capital allocation.skills/maintain-verification-skill/— periodic audit that keeps a project's verification skill and feature map honest.skills/make-bot-ui/— build a custom UI that wakes a Grok Bot over a webhook, optionally exposed on Tailscale.skills/no-comments/— spawn Comment Sicko, fix accepted findings, and encode claimed constraints instead of commenting them.skills/principle-attack-the-premise/— question a shared premise once two or more fixes assuming it have failed the same gate.skills/principle-boundary-discipline/— concentrate validation and error handling at system boundaries; keep internal logic pure.skills/principle-build-the-lever/— build the tool that does or proves the work (codemod, script, generator) instead of doing it by hand.skills/principle-encode-lessons-in-structure/— encode a recurring correction as a lint, flag, or check instead of repeating the instruction.skills/principle-exhaust-the-design-space/— build 2-3 competing prototypes before committing to a novel UI or architectural decision.skills/principle-experience-first/— choose user delight over implementation convenience; fewer polished features over more rough ones.skills/principle-fix-root-causes/— trace each symptom to its root cause and fix it there, not with a guard that silences it.skills/principle-foundational-thinking/— get core data structures right first so downstream code becomes obvious.skills/principle-guard-the-context-window/— route bulk work to subagents; keep summaries, not raw payloads, in the main thread.skills/principle-laziness-protocol/— bias toward deletion and the smallest change that solves the problem.skills/principle-make-operations-idempotent/— design commands and lifecycle steps to converge to the same end state despite crashes and retries.skills/principle-migrate-callers-then-delete-legacy-apis/— migrate callers and delete the old API in the same wave, no compatibility layer.skills/principle-minimize-reader-load/— collapse one-caller wrappers and shrink mutable scope to reduce hidden state a reader must track.skills/principle-model-the-domain/— encode the domain in a structure instead of scattered conditionals.skills/principle-never-block-on-the-human/— proceed on reversible work and let the human course-correct after; reserve confirmation for irreversible actions.skills/principle-outcome-oriented-execution/— converge on the target architecture during a migration instead of preserving throwaway compatibility code.skills/principle-prove-it-works/— verify against the real artifact before declaring done, not a proxy or self-report.skills/principle-redesign-from-first-principles/— redesign as if a new requirement had been foundational from day one, instead of bolting it on.skills/principle-separate-before-serializing-shared-state/— eliminate sharing between concurrent writers first; serialize only when one shared writer is a real invariant.skills/principle-sequence-verifiable-units/— break multi-step work into small units that each end in a verifiable state.skills/principle-subtract-before-you-add/— remove dead code and redundant validators before building on top.skills/principle-test-behavior-not-implementation/— call code the way its users do and assert the result they observe.skills/principle-type-system-discipline/— make illegal states unrepresentable; brand primitives, parse at boundaries, exhaust variants.skills/recall/— reconstruct recent working context from chat history and shared state into a tight current-state brief.skills/reflect/— spawn three parallel review subagents over the active transcript and route learnings to concrete skill edits.skills/show-me-your-work/— keep a reviewable decision trail (TSV log) for long-running or unattended work.skills/swarm/— fan out N parallel workers, drain them, return one report.skills/tdd/— write a failing test then make it pass, when explicitly requested or the bug has an obvious cheap test target.skills/teach/— explain a body of work plainly by runninghowandwhyand weaving the results together.skills/technical-writing/— Diátaxis structure, Google developer style, and STE instruction rules for docs, RFCs, and PR descriptions.skills/typescript-best-practices/— TypeScript best practices for any.ts/.tsxfile.skills/unslop/— cut AI tells from any writing.skills/why/— discover available MCPs and query each evidence source in parallel for a cited read on design rationale and decisions.
- agi-now/buffett-skills — Warren Buffett-style investment analysis via structured stock-analysis templates and reference material.
- mukul975/Anthropic-Cybersecurity-Skills — 818 structured cybersecurity skills mapped to MITRE ATT&CK and NIST CSF 2.0 for security workflows.
- NeoLabHQ/ddd — Bakes Clean Architecture, SOLID, and Domain-Driven Design patterns into the dev workflow via automated rules.