fang2hou

11 mods across 1 repository, 2 stars between them.

create-skill

01

fang2hou/skills

Skill Claude Code

Creates spec-compliant Agent Skills from requirements through validated delivery: intent analysis, domain research, SKILL.md authoring, and with/without-skill evals. Use when creating, drafting, or scaffolding a new Agent Skill.

not rated 2 changed today A 51 tokens original MIT

git-branch-cleanup

02

fang2hou/skills

Skill Claude CodeCodex

Cleans up local Git worktrees and branches whose work is provably merged or absorbed, deleting each risk group only after its own confirmation. Use after merging PRs, after parallel agent worktree sessions, or to prune merged local branches.

not rated 2 changed 20d ago A 54 tokens original MIT

fang2hou/skills

Skill Claude CodeCodex

Applies the standard GitHub repository baseline via the gh CLI: squash-only merges, immutable releases, auto-deleted PR branches, and a Protect Default Branch ruleset. Use for standard GitHub settings, branch protection, or initializing a repository.

not rated 2 changed 20d ago A 55 tokens original MIT

improve-skill

04

fang2hou/skills

Skill Claude Code

Audits and improves existing Agent Skills. Scores 8 quality dimensions, drafts an approval-gated plan, applies changes, and verifies them against a baseline run. Use when reviewing, refactoring, or upgrading a skill or a skill collection.

not rated 2 changed 20d ago A 53 tokens original MIT

review-and-fix

05

fang2hou/skills

Skill Claude Code

Reviews git diff changes with senior-engineer judgment: investigates context, adapts rigor to change type, weighs downstream impact, and guides user-approved fixes. Use when reviewing code changes and deciding whether to fix issues.

not rated 2 changed 20d ago A 47 tokens original MIT

test-design

06

fang2hou/skills

Skill Claude CodeCodex

Designs structured test cases, edge and boundary cases, and coverage gaps from code, specs, or tickets, without writing test code. Use when the user asks for a test plan, test cases, or edge-case analysis before or alongside implementation.

not rated 2 changed 20d ago A 52 tokens original MIT

natural-writing

07

fang2hou/skills

Skill Claude CodeCodex

Rewrites, polishes, drafts, and reviews English, Chinese, and Japanese prose so it reads naturally while keeping facts, terms, and voice. Use to fix AI tone, translationese, stiff or slangy register, or keigo; not for code review or HTML reports.

not rated 2 changed 16d ago A 60 tokens original MIT

agent-report

08

fang2hou/skills

Skill Claude CodeCodex

Default output for any answer with several findings: a desktop HTML report or review page. Use unasked when told to investigate, summarize recent work or commits, compare, audit, or explain a system, or when the user must decide several questions.

not rated 2 changed today A 52 tokens original MIT

figcraft-live

09

fang2hou/skills

Skill Claude CodeCodex

Designs, inspects, renders, and exports in FigCraft (github.com/storytold/figcraft) and .figcraft files via the lazy figcraft-live MCP server (mounts globally, opens one shared app window on ensure) or headless figcraft-cli. Use for FigCraft work.

not rated 2 changed today A 66 tokens original MIT

At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: