my-skills by akitaonrails
akitaonrails my-skills
Fabio Akita's personal set of 24 skills for Claude Code, Codex and opencode: releases, PR audits, security and fact-checking
- 292 ★ on the repository
- Last commit
Extended access
Why this level
- The release skill sets tags and publishes releases: it writes to git and to package registries
- The fact-check and security-audit skills start subagents on other agents' CLIs and read project code
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/akitaonrails-my-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The release skill sets tags and publishes releases: it writes to git and to package registries
- The fact-check and security-audit skills start subagents on other agents' CLIs and read project code
This is third-party code. Review the repository files before installing.
What it does
my-skills is Fabio Akita's public repository of 24 skills. The structure is flat: each folder holds one skill with a `SKILL.md` file. The author links them into Claude Code, Codex, opencode, Kimi and a shared `~/.agents/skills` folder with symlinks, so one edit in the repository reaches every agent. The skills cover a developer's working cycle. `release` derives the version number from the changelog and tags only on green CI. `pr-audit` and `pr-post-audit` review a pull request before and after merge. `security-audit` builds a threat model and looks for vulnerabilities, backdoors and secret leaks. `fact-check` verifies claims in an article against primary sources, and `deepwork` orchestrates large multi-phase tasks. There is also a skill for the `agent-browser` CLI, a `codemap` codebase map and a `humanizer` skill vendored from blader/humanizer. The author warns that the skills are tailored to his machines, subscriptions and workflow and published as examples. He suggests reading them and then asking your own agent to write skills for your needs instead of copying them verbatim.
Who it is for. Developers who write their own skills and want to see how a practitioner organizes them, and people who need reference examples of release, security-audit or fact-check.
Good fit when
- You are writing your own skill for releases or PR audits and want a real model of structure and wording
- You wonder how one skill set is shared across several agents with symlinks: the README table shows paths for Claude Code, Codex, opencode and Kimi
- Before publishing an article you want a fact check: `fact-check` runs checker subagents against primary sources
- Releases ship often and you want the version derived from the changelog and the tag set after green CI
- A third-party pull request has to be audited for vulnerabilities and backdoors
Not a fit when
- You need a ready-made set for someone else's project: by the author's own words the skills are not meant for general use
- The skills would run as is on another machine: they hard-code the author's paths, subscriptions and environment layout
- You want a single install command: the repository has no installer and skills are linked by hand
- You need one skill such as `humanizer`: it is better taken from the original blader/humanizer repository
Example request
Take the release skill from my-skills as a model and write a similar one for my repository: version from the CHANGELOG, tag only after green CI.Limitations
- The author does not intend general use: paths, subscriptions and environment layout are tailored to his machines.
- There is no installer, and skills are linked with symlinks into the agent's skills directory such as `~/.claude/skills` or `~/.codex/skills`.
- Some skills depend on each other: `release` calls `pr-post-audit`, and `fact-check` starts subagents on other agents' CLIs (zcode, claude, codex, opencode, kimi, grok, agy).
- The repository states no license.
How to disable. Skills are connected by symlinks: delete the link from the agent's skills directory (`~/.claude/skills`, `~/.codex/skills`) and the repository itself stays in place. Restart the agent so it stops seeing the skill.
FAQ
How do I connect skills from my-skills to Claude Code?
Clone the repository and for each skill you need create a link with `ln -s ~/Projects/my-skills/<name> ~/.claude/skills/<name>`. Restart the agent so it reads the new `SKILL.md`.
Can I use the skills unchanged?
The author advises against it: the skills carry assumptions about his machines, subscriptions and credential layout. Read the `SKILL.md` you need and adapt it to your project.
What does the fact-check skill do?
It verifies each claim in an article or essay against primary sources online. Cheap checker subagents started on agent CLIs such as claude, codex and opencode do the work, and the skill triggers on requests to check the facts.
Which skills are included?
Among them are `release`, `pr-bump`, `pr-audit`, `pr-post-audit`, `security-audit`, `fact-check`, `deepwork`, `worktrees`, `simplify`, `codemap`, `agent-browser` and `humanizer`. The full list is visible in the repository folders.
Why are some skills missing from the repository?
The author deliberately kept out skills owned by packages, private team and partner assessment sets, opencode plugins and vendor bundles. The list is in the "Deliberately NOT in this repo" section.
Discussion
No questions or reviews yet.
Related
SkillsDozens of mizchi's agent skills: frontend audits, AWS, Cloudflare, SQL, Playwright and OpenTelemetry, installed one by one
The author's personal set: shared agent rules and dozens of skills for GitHub, the browser, macOS, messaging and releases, plus a sync script
A set of independent skills against drift in large codebases: hard-cut, root-cause tracing, lean testing and CI right-sizing.
skills-npm links agent skills shipped in your dependencies in node_modules and listed in the skills field of package.json
Over 1,100 ready-made skills for Claude Code, Codex, Gemini CLI and Cursor, plus an MCP server that picks a skill for the task
A set of skills from real work: npm supply-chain audit, CLI building, Obsidian plugin release and an AGENTS.md hierarchy
README in short
The README opens with the author's warning about the personal nature of the set and a table of connections to seven environments. Then come the list of components deliberately kept out of the repository, an inventory of third-party plugins and skills with update procedures, rules for keeping secrets out and a step-by-step way to add a new skill.
README badge
Original SKILL.md
--- name: deepwork description: High-cost orchestrator workflow for large, high-risk, multi-phase coding efforts with meaningful dependencies and review gates. Do not activate for routine multi-file changes. --- # Deepwork Deepwork is an orchestrator workflow for heavy coding sessions. Use it only when the work is clearly large or high-risk: multiple dependent phases, cross-cutting architectural change, unsafe-to-partially-ship migration, or sustained coordination across several specialist lanes. Do not infer Deepwork merely because a task touches multiple files. Do not use it for trivial edits, quick docs changes, simple bug fixes, or routine bounded features. ## Core Contract When deepwork is active, the orchestrator must manage the work as a scheduler, not as the default implementation worker. ## Setup and Deepwork State - create and maintain a local markdown progress file under `.slim/deepwork/`; - save code/doc deliverables to project paths (e.g. `src/`, `docs/`); reserve `.slim/deepwork/` strictly for progress files; ### Deepwork File Create a task-specific file such as: ```text .slim/deepwork/<short-task-slug>.md ``` Before creating this file—and before planning or delegation—inspect the existing `.gitignore` and `.ignore`. Add only missing entries and do not add duplicates: ```gitignore # .gitignore .slim/deepwork/ ``` ```gitignore # .ignore