agent-skills
Health Warn
- No license — Repository has no license file
- Description — Repository has a description
- Active repo — Last push 0 days ago
- Community trust — 51 GitHub stars
Code Pass
- Code scan — Scanned 5 files during light audit, no dangerous patterns found
Permissions Pass
- Permissions — No dangerous permissions requested
This is a collection of personal agent skills designed for design, writing, and engineering tasks. It provides specialized prompts to help AI assistants review code, create project requirements, write documentation, and manage GitHub issues.
Security Assessment
Overall Risk: Low. The tool executes shell commands via bash scripts and interacts with the filesystem to read plans and write changes. It also makes network requests to interact with the GitHub API (creating issues, opening pull requests, and managing worktrees). However, the automated code scan found no dangerous code patterns, hardcoded secrets, or requests for excessive permissions. The repository is entirely prompt-based and does not harbor hidden malicious payloads.
Quality Assessment
The project is actively maintained, with its most recent push happening today. It enjoys a decent level of community trust with 51 GitHub stars, indicating that other developers have found it useful and reviewed its contents. The primary negative is the complete lack of an open-source license. While you can freely read and run the scripts for personal use, the absence of a license means you technically do not have legal permission to modify, distribute, or use these skills within a commercial product.
Verdict
Safe to use for personal productivity and experimentation, but keep the unlicensed status in mind if you plan to incorporate it into a commercial workflow.
Agent skills I use every day.
Agent Skills
A collection of agent skills for design, engineering, writing, and product, built on the open Agent Skills standard.
Install
Add via skills.sh
npx skills add richtabor/agent-skills
Or as a Claude Code plugin
/plugin marketplace add https://github.com/richtabor/agent-skills
/plugin install rt
Available Skills
Red Pen
A strict editorial reviewer that checks writing, inspired by George Orwell's "Politics and the English Language" rules and Matthew Butterick's Practical Typography.
Accessibility Review
Audits components, pages, and screenshots against WCAG 2.1/2.2 Level AA. Returns prioritized findings (Critical/Warning) with specific fixes rather than generic guidance.
WordPress Add Links
Scans a blog post draft, finds relevant internal and external link opportunities, and weaves them into existing sentences without disrupting the flow.
Create PRD
Plans features interactively. Asks clarifying questions, then generates a detailed PRD document with user stories, acceptance criteria, and technical considerations.
Ralph: Create JSON Issues
Converts a PRD or plan file into right-sized JSON stories for the JSON loop. Reads from .claude/plans/, plans/, or prds/, sizes each story to one context window, and wires up cross-PRD dependencies.
Fresh Eyes
Re-reads code you just wrote as if seeing it for the first time. Catches off-by-one errors, missed edge cases, naming issues, and logic bugs, then fixes them on the spot.
Humanize
Rewrites AI-generated text so it reads like a person wrote it. Applies 30 rules across 6 categories based on Wikipedia's AI writing patterns guide, preserving meaning while removing robotic phrasing.
Markdown Image Alt Text
Reads each image in a markdown file, describes its content, and writes detailed alt text inline. Works with both Obsidian wiki-style (![[image]]) and standard markdown formats.
Motion Design
Evaluates UI animation needs and recommends specific easing curves, durations, and implementation approaches. Every animation gets a job, or it gets cut.
Ralph: GitHub Create Issues
Converts a PRD or plan file into GitHub Issues (parent + sub-issues). Reads from .claude/plans/, plans/, or prds/, sizes stories for single-context execution, and links dependencies via GitHub's native blocking. This is the workflow I use.
Ralph: GitHub Start Loop
Autonomous development loop powered by GitHub Issues. Fetches open PRD issues, spins up git worktrees, implements stories, closes sub-issues on commit, and opens a PR when done.
Ralph: JSON Create Issues
Converts a PRD or plan file into right-sized JSON stories for the JSON loop. Reads from .claude/plans/, plans/, or prds/, sizes each story to one context window, and wires up cross-PRD dependencies. Learn more at richtabor.com/ralph.
Ralph: JSON Start Loop
Autonomous development loop powered by local JSON files. Picks up stories from .claude/plans/, plans/, or prds/, spins up git worktrees, and executes them one by one without manual intervention. Learn more at richtabor.com/ralph.
Review PR
Reviews PR comments from GitHub (Copilot, reviewers), evaluates them against actual code, replies with reasoning, and resolves threads.
Review AGENTS.md
Detects missing AGENTS.md files, creates them using progressive disclosure principles, and symlinks CLAUDE.md. Proactively checks new projects and handles all combinations of existing files.
Skill Review
Validates agent skills against the Agent Skills standard. Audits structure, frontmatter, description quality, progressive disclosure, and anti-patterns, then returns a scored report with actionable fixes.
Technical Writing
Turns the feature you just built into a technical blog post. Analyzes your codebase for implementation details, structures the narrative, and avoids AI-sounding language. Supports WordPress publishing.
X Writing
Turns notes and ideas into X posts that sound like you wrote them. Pulls ideas from a markdown file (X_SOURCE_FILE) or macOS Notes (X_NOTES_APP_SOURCE), identifies the most shareable angle, and drafts posts matched to your voice. Saves approved posts to X_DRAFTS_FILE.
Structure
Each skill lives in its own folder:
skill-name/
├── SKILL.md # Required: Instructions and metadata
├── references/ # Optional: Documentation loaded as needed
├── scripts/ # Optional: Executable code
└── assets/ # Optional: Templates, images, etc.
Credits
The PRD and Ralph skills (ralph-json-create-issues, ralph-json-start-loop) are inspired by snarktank/ralph.
The fresh-eyes skill is inspired by Jeffrey Emanuel.
Reviews (0)
Sign in to leave a review.
Leave a reviewNo results found