agent-skills

Compile a plain-language task into a concise, auditable Codex or Claude Code `/goal`, or explain why a normal prompt fits better. Use when the user asks to draft, formulate, rewrite, tighten, or create a goal for multi-step work that needs a durable objective, transcript-visible proof, constraints, bounded stop conditions, host-aware operation, and risk-based review depth.

Conduct a thorough alignment interview to deeply understand a task before starting work. Use when starting any non-trivial task — take-home exercises, ambiguous problems, design challenges, complex implementations, research questions — anything where shared understanding matters more than speed. Triggers on phrases like "interview me", "let's align on this", "before we start", "kick off this task", "probe me on this", "I have a take-home", "help me think through", "I want to align before we begin", or whenever the user signals they want a deep upfront context-gathering session before diving in. Err strongly toward triggering for any substantive new task — measure twice, cut once. Produces a written kickoff brief that becomes the shared foundation for the work.

Execute architectural refactoring from an assessment document with deterministic, chunked operations and aggressive verification at every step. Use when you have an architectural assessment, clean architecture review, refactoring recommendations, or seam-ripper output and need to actually perform the refactoring safely. Also use when asked to "refactor based on this assessment", "execute these architectural recommendations", "fix architectural drift", "refactor in chunks", or any request to systematically restructure a codebase according to a plan. Designed specifically to prevent the kind of agent drift that causes architectural problems in the first place.

Continuous formal verification of architectural constraints and code quality. Use when asked to verify, audit, or validate codebase integrity. Runs automatically via hooks on every edit (structural) and pre-commit (full). Catches ownership violations, boundary crossings, state machine bugs, and code smells that grep ratchets miss. Triggers: "verify", "formal verify", "check architecture", "audit code quality", "run verification", "/verify", "/verify --bootstrap", "/verify --grade".

Craft a high-quality prompt for a deep research agent (like ChatGPT Deep Research) through adaptive interviewing. Use when the user wants to research something but needs help formulating what to ask — when they say "I need to research X", "help me figure out what to ask about Y", "write a research prompt for Z", "I want to use deep research on...", or when they have a vague research need and want a precise, comprehensive prompt that will get excellent results from a research agent. Also use when the user mentions deep research, ChatGPT research, or preparing a query for an AI research tool.

First-principles simplification analysis for codebases. Methodically inventories what a codebase actually does, then asks whether each piece of complexity earns its keep. Use when asked to "simplify this codebase", "is this overengineered", "how could this be simpler", "reduce complexity", "first principles review", "essential complexity audit", "do we really need all this", or any request to rethink whether the current implementation is the simplest way to achieve its goals. Also useful when a codebase feels harder to work with than it should, when onboarding takes too long, or when changes that seem simple keep ballooning in scope.

This skill should be used when cleaning up codebases that have accumulated dead code, redundant implementations, and orphaned artifacts — especially codebases maintained by coding agents. Triggers on "find dead code", "clean up unused code", "remove redundant code", "prune this codebase", "dead code sweep", "code cleanup", or when a codebase has gone through multiple agent-driven refactors and likely contains overlooked remnants. Systematically identifies cruft, categorizes findings, and removes confirmed dead code with user approval.

Remove LLM-isms and AI writing patterns from text. This skill should be used when editing prose to sound less like AI output — removing overused words, fixing structural tells, and restoring natural human voice. Triggers: "de-slop", "remove AI writing", "humanize this", "sounds too AI", "LLM-isms", "AI slop", or when reviewing text that reads like chatbot output.

Transform a codebase study guide into a polished interactive web experience. This skill should be used when the user has a completed study guide markdown file (from codebase-study-guide or similar) and wants to turn it into an interactive pedagogical app. Triggers on requests like "make this study guide interactive", "turn this into an interactive experience", "visualize this study guide", "create an interactive version", or when a user has a study guide .md file and wants a richer presentation. Produces a Vite-served single-page app with scroll-driven storytelling, interactive architecture diagrams, animated code walkthroughs, and progressive disclosure.

Write changelog entries for open source documentation sites using Keep a Changelog format. Use when asked to "write a changelog", "update the changelog", "add changelog entry", "document recent changes", or after a release/set of changes that should be recorded. Reviews git commits since the last changelog entry and produces a categorized, human-readable entry.

Generate a pedagogically-grounded study guide for learning an unfamiliar codebase. Use when the user wants to onboard onto a codebase, understand a project's architecture, create learning materials for a team, or asks things like "help me learn this codebase", "create an onboarding guide", "I'm new to this project", "how does this system work", "study guide for this repo", or "explain this codebase to me". Produces a structured document that builds understanding from purpose to systems to patterns, using evidence-based learning techniques (elaborative interrogation, concept mapping, threshold concepts, worked examples, progressive disclosure).

Create a self-contained review package of current work for external review by another AI model or human reviewer. Bundles relevant files with a contextual README and instructional prompt. Triggers: "review package", "create review package", "hand off for review", "get a second opinion", "external code review", "cross-model review", "package for review", "prepare code review". Accepts an optional focus area argument to scope the analysis.

Structured development workflow that separates research, planning, and implementation into distinct phases with persistent markdown artifacts. Use when starting any non-trivial feature, refactor, bug investigation, or codebase change. Trigger on: "deep work", "research and plan", "plan before coding", "write a plan", "research this codebase", "don't code yet", "understand then implement", or when the user wants a disciplined approach to a complex task. Also use when the user says "research", "plan", "annotate", "implement the plan", or references research.md/plan.md artifacts.

Make application behavior visible to coding agents by exposing structured logs and telemetry. Use when asked to "add telemetry", "make logs accessible to agents", "add observability", "debug with logs", or when an agent needs to understand runtime behavior but has no way to query logs. Also use when debugging is difficult because there are no structured logs, when agent docs (CLAUDE.md, AGENTS.md) lack instructions for querying application logs, or when setting up logging infrastructure for a new or existing web application.

Research a UI design aesthetic and produce exhaustive, implementation-ready design guidelines for coding agents. Use when the user names an aesthetic (brutalist, glassmorphism, retro-futuristic, Swiss modernist, Apple HIG, neumorphism, minimalism, cyberpunk, Material Design, art deco, vaporwave, etc.) and wants a complete style guide with exact CSS values, color palettes, component states, animations, and typography — detailed enough for a coding agent to faithfully implement the aesthetic with zero ambiguity.

Cold, methodical diagnostician for when you're stuck in agent-assisted app development. Call the fixer when: (1) You're in a loop with a coding agent and things keep getting worse, (2) Your project has accumulated so many agent-generated changes you've lost the thread, (3) Builds are broken and you can't figure out why, (4) You've tried multiple approaches and none are working, (5) You need someone to cut through confusion and give you a clear path forward. Triggers on: "I'm stuck", "nothing is working", "help me fix this", "I'm going in circles", "the agent keeps breaking things", "I've lost track of what's happening", "can you take a look at this mess", or any expression of frustration with agent-assisted development. The fixer does not commiserate — it diagnoses, intervenes, and unblocks.

Generate a smart bootstrap prompt to continue the current conversation in a fresh session. Use when (1) approaching context limits, (2) user says "handoff", "bootstrap", "continue later", "save session", or similar, (3) before closing a session with unfinished work, (4) user wants to resume in a different environment. Outputs a clipboard-ready prompt capturing essential context while minimizing tokens.

Product analytics expert using PostHog MCP. Triggers on requests to understand user behavior, surface insights, create dashboards, analyze funnels, track metrics, set up experiments, or answer questions about product performance. Use when working with PostHog data, discussing analytics strategy, investigating user journeys, retention, conversion, feature adoption, or when asked to help understand what's happening in the product.

Identify non-obvious signals, hidden patterns, and clever correlations in datasets using investigative data analysis techniques. Use when analyzing social media exports, user data, behavioral datasets, or any structured data where deeper insights are desired. Pairs with personality-profiler for enhanced signal extraction. Triggers on requests like "what patterns do you see", "find hidden signals", "correlate these datasets", "what am I missing in this data", "analyze across datasets", "find non-obvious insights", or when users want to go beyond surface-level analysis. Also use proactively when you notice interesting anomalies or correlations during any data analysis task.

Create high-quality animated explainer visuals for essays and blog posts. Use when the user wants to visualize concepts, processes, data, or ideas with interactive web animations. Triggers on requests like "create a visual for", "animate this concept", "make an explainer", "visualize this idea", "diagram this process", "show this data", or when essay content would benefit from visual explanation. Handles abstract concepts (mental models, frameworks), technical processes (algorithms, systems), and data visualization (trends, comparisons). Outputs self-contained HTML/CSS/JS that embeds directly in web content.

CAVEMAN HUNT BAD PROCESS! Me find greedy creature eating fire and rocks. Me bonk them good. Use when tribe say "kill processes", "clean up servers", "save battery", "find resource hogs", "bonk next.js", or "hunt processes". Me bonk known bad creature automatic. Me ask before bonk mystery creature.

Build interactive debugging interfaces that reveal internal system behavior. Use when asked to "help me understand how this works", "show me what's happening", "visualize the state", "build a debug view", "I can't see what's going on", or any request to make opaque system behavior visible. Applies to state machines, data flow, event systems, algorithms, render cycles, animations, CSS calculations, or any mechanism with hidden internals.

Interview-driven blog post drafting for technical product audiences. Use when user wants to write a blog post, article, or essay and needs help developing their thesis, structure, and initial draft. Triggers on "write a blog post", "draft an article", "help me write about X", "blog drafter", or when user has a topic they want to turn into written content. Conducts structured interviews using AskUserQuestion to extract the user's unique insights before generating drafts.

Synthesize outputs from multiple AI models into a comprehensive, verified assessment. Use when: (1) User pastes feedback/analysis from multiple LLMs (Claude, GPT, Gemini, etc.) about code or a project, (2) User wants to consolidate model outputs into a single reliable document, (3) User needs conflicting model claims resolved against actual source code. This skill verifies model claims against the codebase, resolves contradictions with evidence, and produces a more reliable assessment than any single model.

Meta-cognitive decision support that analyzes current context and surfaces intelligent next-step options to the user. Use this skill when: (1) User explicitly invokes /checkpoint, (2) Significant work has been completed and a checkpoint is valuable, (3) Uncertainty or ambiguity exists about requirements or approach, (4) Task complexity has expanded beyond initial scope, (5) Before finalizing or committing to ensure nothing is missed. This skill pauses execution, assesses the situation holistically, and presents 2-5 contextually-appropriate options via AskUserQuestion, with a recommended option and rationale.

Perform comprehensive, deep analysis of a system and its subsystems to identify bugs, race conditions, stale documentation, dead code, and correctness issues. Use when asked to "audit this system", "exhaustive analysis of X", "analyze for correctness", "root out issues in...", "deep dive into...", "verify this code is correct", "find bugs in...", or when reviewing agent-written code for production readiness. Automatically decomposes systems into subsystems, applies appropriate analysis checklists, and produces structured findings with severity classification.

Build a retrieval-optimized knowledge layer over agent documentation in dotfiles (.claude, .codex, .cursor, .aider). Use when asked to "optimize docs", "improve agent knowledge", "make docs more efficient", or when documentation has accumulated and retrieval feels inefficient. Generates a manifest mapping task-contexts to knowledge chunks, optimizes information density, and creates compiled artifacts for efficient agent consumption.

Expertise in architecting, implementing, reviewing, and debugging hierarchical matching systems. Use when working with: (1) Two-sided matching (Gale-Shapley, hospital-resident, student-school), (2) Assignment/optimization problems (Hungarian algorithm, bipartite matching), (3) Multi-level hierarchy matching (org charts, taxonomies, nested categories), (4) Entity resolution and record linkage across hierarchies. Triggers: debugging match quality issues, reviewing matching algorithms, translating business requirements into constraints, validating match correctness, architecting new matching systems, fixing unstable matches, resolving constraint violations, diagnosing preference misalignment.

Facilitate methodical review of proposals (technical designs, product specs, feature requests). Use when asked to "review this proposal", "give feedback on this doc", "help me review this RFC", or when presented with a document that needs structured feedback. Handles markdown files, GitHub gists/issues/PRs, and other text formats. Chunks proposals intelligently, predicts reviewer reactions, and produces feedback adapted to the proposal's format.

Enter todo recording mode to capture ideas without acting on them. Use when the user says "record todos", "let's capture some todos", "brainstorm mode", or wants to dump ideas without immediate execution. Captures thoughts to .claude/todos/, then organizes and prioritizes on exit.

Compile an agent-optimized changelog by cross-referencing git history with plans and documentation. Use when asked to "update changelog", "compile history", "document project evolution", or proactively after major milestones, architectural changes, or when stale/deprecated information is detected that could confuse coding agents.

Guide users step-by-step through manually testing whatever is currently being worked on. Use when asked to "test this", "verify it works", "let's test", "manual testing", "QA this", "check if it works", or after implementing a feature that needs verification before proceeding.

Assess a codebase's readiness for autonomous agent development and provide tailored recommendations. Use when asked to evaluate how well a project supports unattended agent execution, assess development practices for agent autonomy, audit infrastructure for agent reliability, or improve a codebase for autonomous agent workflows. Triggers on requests like "assess this project for agent readiness", "how autonomous-ready is this codebase", "evaluate agent infrastructure", or "improve development practices for agents".

Write clear, plain-spoken code comments and documentation that lives alongside the code. Use when writing or reviewing code that needs inline documentation like file headers, function docs, architectural decisions, or explanatory comments. Works well for both human readers and AI coding assistants who see one file at a time.

Robust Rust patterns for file-backed data, parsing, persistence, FFI boundaries, and system integration. Use when writing Rust that handles file formats, subprocess integration, PID/process management, Serde serialization, or UniFFI boundaries. Covers UTF-8 safety, atomic writes, state machines, and defensive error handling.

Design intuitive, meaningful interactions grounded in user goals and cognitive principles. Use when designing component behaviors, user flows, feedback systems, error handling, loading states, transitions, accessibility, keyboard navigation, touch/gesture interactions, or when evaluating interaction quality. Also use for modal vs modeless decisions, direct manipulation patterns, input device considerations, emotional/dramatic aspects of UX, or when asked about making interfaces feel responsive, humane, and goal-directed.

Analyze recent conversation context and capture learnings to project knowledge files (for project-specific insights) or skills/commands/subagents (for cross-project patterns). Use when the user asks to "capture this learning", "update the docs with this", "remember this for next time", "document this issue", "add this to CLAUDE.md", "save this knowledge", or "update project knowledge". Also triggers after resolving build/setup issues, discovering non-obvious patterns, or completing debugging sessions with valuable insights.

Apply Model-First Reasoning (MFR) to code generation tasks. Use when the user requests "model-first", "MFR", "formal modeling before coding", "model then implement", or when tasks involve complex logic, state machines, constraint systems, or any implementation requiring formal correctness guarantees. Enforces strict separation between modeling and implementation phases.

Create visual parameter tuning panels for iterative adjustment of animations, layouts, colors, typography, physics, or any numeric/visual values. Use when the user asks to "create a tuning panel", "add parameter controls", "build a debug panel", "tweak parameters visually", "fine-tune values", "dial in the settings", or "adjust parameters interactively". Also triggers on mentions of "leva", "dat.GUI", or "tweakpane".

Expert Unix and macOS systems engineer for shell scripting, system administration, command-line tools, launchd, Homebrew, networking, and low-level system tasks. Use when the user asks about Unix commands, shell scripts, macOS system configuration, process management, or troubleshooting system issues.