Skills

Source-of-truth repository for agent skills. A skill is a directory with a SKILL.md (YAML frontmatter + instructions), following the Agent Skills format.

Skills live in skills/ and are synced additively to two deployment targets:

Target	Folder	Script
OpenCode	%USERPROFILE%\.config\opencode\skills	mirror-skills-opencode.ps1
Cross-harness (.agents)	%USERPROFILE%\.agents\skills	mirror-skills-agents.ps1

The .agents folder is the de-facto global skills location read natively by Codex CLI, Cursor, Gemini CLI, VS Code/Copilot, and OpenCode (compat). Claude Code does not read it (it uses ~/.claude/skills).

Quick Start

# 1. Clone
git clone <this-repo-url> skills
cd skills

# 2. Mirror all skills to both targets
.\mirror-skills-opencode.ps1   # -> %USERPROFILE%\.config\opencode\skills
.\mirror-skills-agents.ps1     # -> %USERPROFILE%\.agents\skills

Layout

skills/
├── AGENTS.md                  # Rules for agents working in this repo
├── README.md
├── mirror-skills-opencode.ps1 # Additive sync: repo -> OpenCode skills folder
├── mirror-skills-agents.ps1   # Additive sync: repo -> .agents skills folder
├── skills-maintenance/        # Meta skill: how to maintain this repo
└── skills/
    └── <skill-name>/
        ├── SKILL.md           # Required
        └── ...                # Optional scripts / references / assets

The meta skill skills-maintenance/ lives at the repo root (not in skills/) because it maintains this repo rather than serving end users. It is mirrored to both targets like any other skill.

The Mirror Scripts

Both scripts are additive and never destructive:

• Add skills that exist in the repo but not in the target.
• Update files of existing skills when they changed in the repo.
• Never remove skills or files that exist only in the target folder.
• Clean copied files of NTFS alternate data streams (e.g. the Zone.Identifier "Mark of the Web" that Windows attaches to downloaded files), so that junk never propagates into the target folder.

.\mirror-skills-opencode.ps1              # sync to %USERPROFILE%\.config\opencode\skills
.\mirror-skills-agents.ps1                # sync to %USERPROFILE%\.agents\skills
.\mirror-skills-opencode.ps1 -WhatIf      # preview changes only
.\mirror-skills-agents.ps1 -TargetPath X  # sync to a custom folder

Each script prints a summary of added / updated / unchanged / skipped skills.

Workflow

1. Create or edit a skill in skills/ (see skills-maintenance/SKILL.md for the skill format and full maintenance guide).
2. Update the "Skills" list below if you added, removed, renamed, or re-described a skill.
3. Commit and push:

   git add -A
   git commit -m "add: my-skill"
   git push origin master

4. Sync to both targets:

   .\mirror-skills-opencode.ps1
   .\mirror-skills-agents.ps1

Removing a skill

The mirrors never delete. To retire a skill: delete its directory from skills/, remove it from the list below, commit and push, then manually remove the folder from each deployment target.

Skills

Skill	Description
agent-browser	Browser automation CLI for AI agents: navigate pages, fill forms, click, screenshot, scrape data, and test web apps and Electron apps.
author-voice	Build, calibrate, and apply a reusable author voice profile so editing and drafting preserve a specific writer's identity without imitating accidental typos.
business-model-canvas	Design and analyze business models using the Business Model Canvas framework.
ce-brainstorm	Explore vague or ambitious ideas into a right-sized, requirements-only plan before building.
clonedeps	Clone dependency source code into a local workspace so agents can inspect library internals.
code-architecture-tailwind-v4-best-practices	Tailwind CSS v4 patterns for buttons and components (CVA/tailwind-variants, CSS-first config).
code-architecture-wrong-abstraction	When to abstract vs. duplicate code; DRY/WET decisions and refactoring abstractions.
codemap	Generate comprehensive hierarchical codemaps for unfamiliar repositories.
cognitive-biases	Apply cognitive bias knowledge to product design and decision-making.
cognitive-fluency-psychology	Use cognitive fluency principles to improve clarity, trust, and conversion.
cognitive-load	Design experiences that optimize mental resources using Cognitive Load Theory.
css-container-queries	CSS container queries for component-based responsive design.
curiosity-gap	Create engagement through strategic information gaps that drive user action.
customer-research	Conduct, analyze, and synthesize customer research (interviews, surveys, reviews, community mining).
deepwork	High-cost orchestrator workflow for large, high-risk, multi-phase coding efforts.
find-skills	Help users discover and install agent skills.
five-whys	Root cause analysis using the Five Whys technique.
fogg-behavior-model	Design behavior change using the B=MAP framework.
graph-thinking	Graph-based thinking for mapping dependencies and solving problems non-linearly.
halo-effect-psychology	Apply the halo effect in product design and UX.
hicks-law	Optimize decision speed by managing the quantity of choices.
hypothesis-tree	Structure complex questions into testable hypotheses.
image	Create, generate, edit, and optimize marketing images with AI tools.
jobs-to-be-done	Understand customer motivations through jobs-to-be-done theory.
loop-engineering	Loop engineering runtime: Grill + Monitor.
loss-aversion-psychology	Leverage loss aversion in product design and messaging.
make-interfaces-feel-better	Design engineering principles for making interfaces feel polished.
making-product-decisions	Framework for structured product decision-making and tradeoff analysis.
marketing-psychology	Apply psychological principles and behavioral science to marketing.
media-recommender	Build and apply profile-driven cross-media recommendations while separating personal fit, execution confidence, and medium/platform path.
motion	Animations with Motion Vue (motion-v) for Vue 3/Nuxt.
naming-cheatsheet	Language-agnostic naming conventions using the A/HC/LC pattern.
nextjs-image-art-direction	Art direction for Next.js images with getImageProps() (per-viewport images).
nuxt	Nuxt 4+ patterns: server routes, file-based routing, middleware, h3 v1, nitropack v2.
oh-my-opencode-slim	Configure and improve oh-my-opencode-slim (agents, models, prompts, skills, MCPs, presets).
pest-analysis	PEST (Political, Economic, Social, Technological) analysis for strategic planning.
pnpm	Manage Node.js dependencies with pnpm: workspaces, catalogs, overrides, patches, CI.
progressive-disclosure	Reduce complexity by revealing information progressively.
project-structure	React/Next.js/TypeScript project organization using feature-based architecture.
react-key-prop	Proper usage of the key prop in React lists.
react-use-callback	Proper usage of the useCallback hook in React.
react-use-client-boundary	Proper usage of the "use client" directive in React/Next.js.
react-use-state	Proper usage of the React useState hook.
reflect	Review recent work, find repeated workflow patterns, suggest reusable skills, agents, and playbooks.
release-smoke-test	Test an oh-my-opencode-slim release candidate or bugfix before publishing.
repository-docs-governance	Audit and restructure a repository's AGENTS.md hierarchy, codemaps, and documentation.
self-initiated-triggers	Design internal triggers for sustained user engagement.
simplify	Simplify and refine recently modified code for clarity and consistency.
social-proof-psychology	Leverage social proof to build trust and influence user behavior.
status-quo-bias	Design for users' preference for the current state over change.
systematic-debugging	Systematic debugging before proposing fixes for any bug or test failure.
theme-epic-story	Structure product work hierarchically using themes, epics, and stories.
trust-psychology	Build trust signals that reduce perceived risk and enable user action.
ts-library	Authoring TypeScript libraries: setup, exports, build tooling, API design, publishing.
tsdown	Bundling TypeScript libraries with tsdown: dual ESM/CJS, .d.ts, validation.
typescript-advanced-types	TypeScript advanced type system: generics, conditional/mapped types, template literals.
typescript-best-practices	TypeScript best practices for type safety, organization, and maintainability.
typescript-interface-vs-type	When to use interface vs. type in TypeScript.
typescript-satisfies-operator	Proper usage of TypeScript's satisfies operator vs. type annotations.
user-story-fundamentals	Capture requirements as structured user stories with acceptance criteria.
verification-before-completion	Run verification commands and confirm output before claiming work is complete.
verification-planning	Plan project-specific verification evidence for non-trivial coding work.
video	Create, generate, and produce video content with AI tools or programmatic frameworks.
visual-cues-cta-psychology	Design effective CTAs using visual attention and gaze psychology.
vite	Version-aware Vite build, configuration, plugin, SSR, and migration guidance.
vitest	Unit/integration tests for Vite projects with vitest.
vue	Vue 3 patterns: Composition API, props/emits, VueUse, reactive destructuring.
web-perf	Diagnose and improve real-world web performance (Vue 3, Vite, Nuxt, Electron, Tauri).
worktrees	Manage Git worktrees as isolated coding lanes for complex or parallel work.
writing-web-documentation	Write and organize developer-facing documentation for web software projects.

Notes

• A skill is any directory under skills/ containing a SKILL.md. Directories without one are skipped by the mirrors with a warning.
• Don't edit skills directly in a deployment target - the repo is the source of truth and the next mirror run will overwrite repo-tracked files.