Ask HN: How do you manage skills files? (self)

Skills are mostly snake oil, the way people use them (the aspiration to download kung foo from a celebrity).

There was a time when maybe it mattered (last year), but with good repos and good prompts today's agents can find exactly what they need without any skills.

"Skills" as developer macros can be useful, but at most those are things shared with the team (in the repo), not something you download from the internet. If you have so many skills that you feel the need to manage them, that's a code smell.

- I don't find skills, I create them

- Keep them organised in software repos that you install with symlinks for all coding harnesses that you have. Progressive disclosure based on the frontmatter does the rest.

- I make sure they work with AI evals. Think of them like integration tests to prove behaviour. They're useful to optimize your flows. I try to make my skills be mostly a translation between natural language and good small fast tools that they call.

- I change them as a new problem arises. Not just because.

Skills can't be eaten by model capabilities if skills represent a workflow that is custom to my team or my person.

I wrote about a good mental model in the past:

https://alexhans.github.io/posts/series/evals/building-agent...

Never understood the “skills” part tbh. Thse models are being trained in trillions of texts. Adding something small and particular about my codebase doesn’t really move the needle

Surprised it has not been mentioned, but I think relying on https://github.com/vercel-labs/skills is sound. It handles global installs for a wide range of coding agents. If I was working on an internal only skill I'd probably still use the same foundation.

Personally I have found "meta-prompting" to be much more useful than any pre-canned set of skills. Simply ask the AI (inside of a workspace already set up):

Create a standalone prompt to <xyz>

The latest AIs will print out a long prompt with all of the assumptions, tools, and general files it plans to use. Review that, and then run the whole prompt in a new context.

I don't use any skills, what kinds of skills are people finding most useful?

For general tasks, the model seems perfectly capable of figuring out things itself, for project or environment specific tasks, I just put that information in the readme or agents.md file.

Last week I had to reuse homemade skills on different project. I very much liked the AI proposed solution and it works quite well: ship as a plugin and add your git repo as a marketplace.

The installation is effortless and I don't have to mess with symlinks as I may be working with same codebase on different platforms which would make things.. different.

  codex plugin marketplace add "https://path-to-my-git-repo"
  codex plugin add agent-tools@mycompany

  claude plugin marketplace add "https://path-to-my-git-repo"
  claude plugin install agent-tools@mycompany

Let the AI generate .json files for marketplace.

Haven't got to these bits yet, but I'm sure they will work as easy as install does.

  claude plugin marketplace update mycompany
  claude plugin update agent-tools@mycompany

I don’t seem much point to intentionally curating a set of skills, and specifically invoking them by name, only to watch the ai skip them all and do better by just reading code and internal/external websites.

I've got a skill repository on github, and I got a hermes automation to sync skills repo - this hermes automation is on every device. Any skills I make on the fly - my global AGENTS.md has instructions to update the skills repo path and place them appropriately in there.

for skills related to specific cli tools, i just wrote a standard for this! it's obviously not widely used yet, but since mise will support installing the skills alongside the tool, i suspect it will have decent adoption

i used to be a bit bearish on skills—thinking that llms should just use --help, but i've come around on that. i think skills are a great way to describe higher level workflows that use multiple commands.

https://jdx.dev/posts/2026-09-05-introducing-packslip/

When I work on a new problem and the agent struggles with it, I will ask the agent to distill the knowledge and experiences it gained during the session into a skill file. Then I review and publish it depends on the assistant system. I find this is an effective way for the agents to learn new skills, both from its own discoveries but also from my steering and the mistakes it made.

If you work in a niche or on special problems, this template could be useful.

Nowadays I often create skills myself (or with the aid of coding assisants) for any repeating tasks. For example, I use my own skill to make Claude CLI send the worktree to codex cli for the review, then read the verdict and make changes

We have an ongoing discussion and some fans of https://microsoft.github.io/apm/

Because it's from Microsoft and sounds sufficiently enterprisey probably.

The internet has broken me. Whenever I see a question like this i now automatically expect it to be some marketing attempt. There will be a product/service/blog post somewhere in the comments.

Here's how I do it:

- Explanation: https://www.minid.net/2026/7/14/how-to-automatise-with-ai

- Git source: https://github.com/meerita/monorepo-nextjs-golang-rust-pytho...

I have created a little system for this, placed directly in the ~.<youragent>/skills folder.

I have a configuration file of marketplaces and other skills to fetch, it can look like. I have my own marketplaces as well, including ones from my company. I use vercel's tool for managing skills with npx, but to easily handle specifically _which_ skills to fetch, the config file is set up as follows:

  SOURCES = {
    'some-marketplace-name': [
        'some-skill',
        'another-skill',
        'yet-another',
    ],
    'https://designsystem.yourcompany.com': [], (empty list: fetch all skills)
  }

from there I simply run "skills.py" (a single helper) to clean/fetch updated versions of the skills.

I don't use skills unless I have something specific to tell the agent. For example, if I want the agent to use Tailwind V3 instead of V4, I'll have a skill for that. Or, if I want the agent to always use the repository pattern for database access, I'll create a skill for that.

I don't need to manage skills files because I have so few of them and they're only a couple lines long.

5 Stages using local GIT (no remote, don't need it) to prevent preloading in the prompt:

1. A single Skill finder skill, loaded in the prompt, prevents having to import all the summaries in the prompt the harness would add. Uses git's own search.

2. Private repo, per agent, contains main (production) and draft-<name of skill> branches.

3. Shared repo, like 2, but general access for all group agents.

4. Fallback mode, search the harness for skills using the harness mechanism when a relevant skill cannot be found.

5. Skill audit cron. Identify junk skills / drafts that have never changed / not in any recent sessions history, and categorise monthly for me to decide.

This means it's compatible with existing skill folders, removal of git and the finder skill is non destructive and critically debloats the prompt of skills that aren't used and lazy loads them when needed.

I commit them to git(so complete team leverages them)., each repo has kind of different skills and the skills are the ones which I update at least twice a week. I’ve skills on how to add instrumentation , debug, code, code review, tech design review etc. I found most of the skills I find on skills.sh are not very useful for me., but I browse occasionally to get some inspiration. One more paradigm I’m seeing good results on adding new skills is ‘how to do X’, for instance ‘how to add logs’., “how to review code” etc., if i’m not able to frame it that way I don’t think it’s a good use case for me to add that skill to the llm arsenal.

Another thing i discovered is less is more (in case of skills as well)., don’t add lots of skills., keep them very handful - I’ve got 9 skills so far (many people have 100s installed from marketplaces and plugins)

To the extent that skills are contextual guidance (for this author, this project, etc) and not just (raw) capabilities they are unlikely to be eaten by models.

I maintain all my skill files in a central location (like dotfile management) and have guix home sync it to the skill folders of various harnesses that I'm playing with (codex, pi, antigravity, Claude Code, Deepseek harness, etc). They're set up to be bidirectional links rather than read-only like the default configuration, so I can keep editing them / adding to the corpus from any harness.

This works well for skills since all harnesses expect the same format, but is more annoying for other features.

EDIT: This is actually an example of a potentially useful skill. You might choose to manage your skills slightly differently. All you need to do is write a skill-management skill for your agents to be able to wire things up correctly / access them for edits.

Some other nifty skills/plugins in my experience: render latex equations, cetz diagrams inline, jujutsu, guix, code reviewer, writing feedback.

I manage them as part of my dotfiles using chezmoi. A `.agents/skills/` directory + a symlink to there from `.claude/skills/`.

> Do you keep improving them over time?

In my global AGENTS.md I have a note to agents to explain any frustrations they had doing a task, and to suggest any skill/tool/AGENTS.md improvements. I am trying to keep AGENTS.md files small but still finding the balance.

I'm on mobile but the first skill I made was a skill improvement skill. This is basically stating that if the AI struggles in another skill but finds a way that the skill it used needs reworking and it needs to do so.

There is a rule to always use this skill and then track notes in a version file. Then back it up in a share folder or external drive.

Skills have made my tools immensely better, cheaper to use and faster. I've also added to it that it should write scripts it can just use in the future to do tasks like query information it needs to answer questions.

I wish there was a better way to share these over a team but I haven't taken that time yet.

Skills that are so generic that you can find them on the internet, and which you think can be replaced by model improvements, are useless, possibly even harmful, considering how much the models get clingy to the context. Useful skills describe workflows specific to your project, and they can live in the project repo for everyone to use and improve.

The main problem I encountered around this is that skills need to be edited across projects and across team members in a controlled way. Git is of course required for this but is not enough so I built a tool to do just that:

https://github.com/genged/capshelf

Using capshelf I manage my skills across projects. When I start a new project I can just:

$ capshelf add security-review

From the skill repo.

And if I create a new skill I can promote it to the repo so everyone can install it:

$ capshelf promote security-review

It pins the skill content hash so there are no unexpected edits that can break your flow. It also supports MCP configs and agent configs.

Skills live in two source-of-truth git repos (private and public). Agents edit skills by my request, and syncs to all coding agents ~/.claude/skills/, ~/.codex/skills, ~/.pi/agent/skills, ~/.config/opencode/skills etc. with agent written sync-agent-skill script. script ensures that no local changes was made in-place.