agent-skills

A personal collection of Claude Code agent skills and MCP servers, can be used with other agents too.

Available Skills

Skill	Command	Description
git-workflow	/git-workflow	Guided Git workflow: branching, commits, PRs, and conflict resolution
code-review	/code-review	Thorough code review covering correctness, security, performance, and style
english-humanizer	/humanize	Humanize the text
grill-master	/grill, /grill-master, /grill-me	Get relentlessly interviewed about a topic, plan or design until every branch of the decision tree is resolved.
spec-writer	/spec, /spec-writer	Create a PRD through user interview, codebase exploration, and module design, then submit as a GitHub issue. Use when user wants to write a PRD, create a product requirements document, or plan a new feature.
slice-the-spec	/slice, /slice-the-spec	Break a PRD into independently-grabbable GitHub issues using vertical slices.
incremental-tdd	/tdd, /incremental-tdd	Test-driven development with a red-green-refactor loop. Builds features or fixes bugs one vertical slice at a time.
deep-codebase-audit	/deep-codebase-audit, /deep-audit, /codebase-audit	Identify architectural friction in a codebase and refactor shallow, tightly coupled modules into deeper, more cohesive ones—improving testability, structure, and making it easier (and safer) for AI agents to understand and modify.
spec-to-plan	/spec-plan, /spec-to-plan	Turn a PRD into a multi-phase implementation plan using tracer-bullet vertical slices.
domain-glossary	/domain-glossary, /glossary	Turn an ongoing conversation into a DDD-style ubiquitous language document. It scans for domain terms, resolves ambiguities, proposes canonical names, and writes a living glossary to DOMAIN_GLOSSARY.md in the working directory.
script-writer	/script-writer, /write-a-script	Script-writer that drafts presentations, essays, emails, and slides using only the cognitive and persuasive heuristics from Patrick Winston's "How to Speak" lecture.

MCP Servers

Server	Description
ast-editor	AST-targeted code editing MCP server with 28 surgical tools: functions, classes, methods, fields, parameters, imports, parametrized tools for leading comments (edit_leading_comment(op=...)), sibling inserts (insert_sibling), body inserts (insert_in_body(at/after/before)), and tiered reads (read_symbol(depth=full/interface/signature)). In-body snippet editing (replace_in_body, delete_in_body), dict/list editing across JSON/YAML/TOML and Python/JS/TS literals. Dotted targets descend into closures (Go stdioCmd.RunE, TS app.handler). Built on tree-sitter. Supports 11 languages: Python, JavaScript, TypeScript, C, C++, Ruby, Go, Java, JSON, YAML, and TOML.
notebook-editor	Jupyter notebook editing MCP server with 23 cell-level tools: notebook creation (create_notebook with valid nbformat schema), cell structure (add/delete/move/split/merge), content editing (replace/prepend/append), outputs & metadata management, notebook-wide symbol discovery, and kernel-agnostic execution (execute_cell, execute_all_cells, restart/interrupt/shutdown kernel — works with any installed Jupyter kernel: python3, ir, julia-1.x, ...). Complements ast-editor — use this for .ipynb files, ast-editor for everything else.
rmd-editor	R Markdown editing MCP server with 24 chunk- and prose-level tools: document creation (create_rmd), cell structure (add/delete/move/split/merge), content editing (replace/prepend/append), chunk options (get/set/remove with a special label key), YAML frontmatter (get/set/remove), text search, and R kernel execution via IRkernel (outputs returned inline, never persisted — R Markdown regenerates on render). Complements notebook-editor — use this for .rmd files, notebook-editor for .ipynb.

Skills installation via skills.sh

Install multiple skills at once:

npx skills@latest add kambleakash0/agent-skills

or, install individual skills:

npx skills@latest add kambleakash0/agent-skills/skills/git-workflow
npx skills@latest add kambleakash0/agent-skills/skills/code-review
npx skills@latest add kambleakash0/agent-skills/skills/english-humanizer
npx skills@latest add kambleakash0/agent-skills/skills/grill-master
npx skills@latest add kambleakash0/agent-skills/skills/spec-writer
npx skills@latest add kambleakash0/agent-skills/skills/slice-the-spec
npx skills@latest add kambleakash0/agent-skills/skills/incremental-tdd
npx skills@latest add kambleakash0/agent-skills/skills/deep-codebase-audit
npx skills@latest add kambleakash0/agent-skills/skills/spec-to-plan
npx skills@latest add kambleakash0/agent-skills/skills/domain-glossary
npx skills@latest add kambleakash0/agent-skills/skills/script-writer

Skill Format

Each skill is a Markdown file with a YAML front-matter block followed by the prompt body, following the skills.sh spec:

---
name: skill-name
description: >
  What the skill does and when to trigger it.
  TRIGGER when the user writes /skill-name.
metadata:
  author: kambleakash0
  version: x.y.z
triggers:
  - /trigger-1
  - /trigger-2
  - ...
---

Skill prompt body goes here…

Adding New Skills

1. Create a new folder in the root of skills/. The folder name must strictly match your skill's name in the YAML frontmatter.
2. Add a SKILL.md file in the root of the new folder. Ensure it contains a name (max 64 chars) and description (non-empty, max 1024 chars) inside the YAML frontmatter.
3. If needed, add a resources/ folder in the root of the new folder and populate it with the necessary files.
4. Add a row to the Available Skills table in this README.
5. Test it locally by placing the folder from skills/ in ~/.claude/skills/ or ~/.agents/skills/.

MCP Servers' installation

Refer to their respective READMEs for installation instructions.