Skip to content

knightsrule/claude-skills-public

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Claude Code Skills

A collection of reusable Claude Code skills for product management and SRE workflows.

What Are Skills?

Skills are markdown instruction files that teach Claude Code how to perform specific tasks. When invoked, Claude loads the skill and follows its instructions — like a runbook it executes autonomously. Skills can be triggered via /slash-commands or automatically when Claude detects a matching request.

Skills in This Repo

Product

Skill Slash Command Description
incubate /incubate Product incubation and PRD creation through collaborative discovery. Takes a rough idea and turns it into a validated product concept with market research, competitive analysis, and go-to-market strategy.
feature-design /feature-design Feature design conversation that produces a Feature Requirement Document (FRD). Focuses on the what and why before jumping into implementation.
feature-decompose /feature-decompose Breaks a feature (from an FRD or idea) into implementable INVEST stories and creates them as trackable tasks.
prioritize /prioritize RICE-scores stories in the backlog and recommends priorities with a ranked table and approval workflow.

Typical flow: /incubate (PRD) → /feature-design (FRD) → /feature-decompose (stories) → /prioritize (ranking)

SRE

Skill Slash Command Description
fastapi-server /fastapi-server Scaffolds a production-ready FastAPI server from templates with logging, CORS, request tracking, error handling, and health checks.
infosec-check /infosec-check Pre-production security audit — checks DNS/email (SPF, DMARC, DKIM), TLS, HTTP security headers, exposed endpoints, rate limiting, and optionally Cloudflare configuration.
promote-to-main /promote-to-main Release review across multiple repos — diffs dev vs. main, performs code review, detects new env vars, recommends version bumps, and gives a GO / NO-GO verdict.

Installation

Option 1: Personal Skills (Available in All Projects)

Copy individual skill files into your personal Claude skills directory:

# Clone the repo
git clone https://github.com/knightsrule/claude-skills-public.git

# Copy a single-file skill
cp claude-skills/Product/incubate-SKILL.md ~/.claude/skills/incubate/SKILL.md

# Copy a directory-based skill (with assets)
cp -r claude-skills/SRE/fastapi-server ~/.claude/skills/fastapi-server

Each skill must live in its own directory under ~/.claude/skills/ with the entrypoint named SKILL.md:

~/.claude/skills/
├── incubate/
│   └── SKILL.md
├── feature-design/
│   └── SKILL.md
├── fastapi-server/
│   ├── SKILL.md
│   ├── assets/
│   ├── references/
│   └── scripts/
└── ...

Option 2: Project-Level Skills (Shared with Your Team)

Copy skills into your project's .claude/skills/ directory so anyone who clones the repo gets them:

mkdir -p .claude/skills

# Copy a skill into your project
cp claude-skills/SRE/infosec-check-SKILL.md .claude/skills/infosec-check/SKILL.md

Option 3: Install All Skills at Once

git clone https://github.com/knightsrule/claude-skills-public.git

# Install all single-file skills
for f in claude-skills/**/*-SKILL.md; do
  name=$(basename "$f" -SKILL.md)
  mkdir -p ~/.claude/skills/"$name"
  cp "$f" ~/.claude/skills/"$name"/SKILL.md
done

# Install directory-based skills
cp -r claude-skills/SRE/fastapi-server ~/.claude/skills/fastapi-server

Verifying Installation

Open Claude Code and type / — your installed skills should appear in the autocomplete list.

Usage

Invoke any skill by typing its slash command in a Claude Code session:

/incubate I have an idea for a developer productivity tool that...

/feature-design add dark mode toggle

/infosec-check https://app.example.com https://api.example.com

/promote-to-main

/fastapi-server my-api --description "User management API"

Some skills (like feature-design and incubate) are conversational — they'll ask questions before producing output. Others (like infosec-check and promote-to-main) run autonomously and produce a report.

Skill Anatomy

Each skill is a markdown file with YAML frontmatter:

---
name: skill-name
description: When and how Claude should use this skill
---

Instructions that Claude follows when the skill is invoked...

The description field is always in Claude's context and determines when the skill auto-triggers. The full instructions load only when invoked.

Frontmatter Options

Field Purpose
name Slash command name (defaults to directory name)
description When to trigger — Claude reads this to decide relevance
disable-model-invocation true = only manual / invocation, no auto-trigger
allowed-tools Tools Claude can use without asking permission
model Override the model for this skill (e.g., opus)
context fork = run in isolated subagent context

Dependencies

Some skills rely on external tools:

Beads (Task Management)

The feature-decompose and prioritize skills use Beads (bd) for task tracking. Beads is a lightweight, git-native task manager that stores tasks as files in your repo.

Install beads before using these skills:

pip install beads-cli

Initialize in your project:

bd init            # Standard — commits beads files to the repo
bd init --stealth  # Personal — keeps beads files out of git

Skills that use beads: feature-decompose, prioritize

If beads isn't installed, these skills will let you know and suggest setup steps. The other skills (incubate, feature-design, infosec-check, fastapi-server, promote-to-main) have no external dependencies beyond Claude Code itself.

Contributing

To add a new skill:

  1. Create a SKILL.md file (or a directory with SKILL.md + supporting files)
  2. Place it under the appropriate category (Product/, SRE/, or create a new one)
  3. Use the naming convention <skill-name>-SKILL.md for single-file skills
  4. Include clear name and description frontmatter
  5. Update this README with the new skill

License

MIT

About

Reusable Claude Code skills for product management and SRE workflows

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages