Maggie Appleton @maggieappleton.com · 27/04/2026How is everyone managing their agent SKILL.md files? Is it just chaos? Global skills, repo-specific skills, keeping them in sync between machines, figuring out which ones you have installed, authoring new ones. What are we doing? 30740
owais @desertthunder.dev · 27/04/2026I don't use skills conventionally, but I draft < 150 line framework guides with patterns and documentation links that I add to prompts. Skills pollute context and I'm too pedantic to let a model run wild 120
Angadh Nanjangud @angadh.com · 27/04/2026My suspicion (from a call for the with Michael Cohen at the “Build with Opus 4.7” hackathon this weekend) is that this is what is “solved“ by Managed Agents. 100
Jay Gruber @jays.name · 27/04/2026I've got a personal repo where I manage user-scoped skills, mcp.json. I use github.com/vercel-labs/... for "install skill", either from local filepath or from git URL. I have a package.json-like file listing each skill I install. I have a script that parses that file and syncs them all.github.comGitHub - vercel-labs/skills: The open agent skills tool - npx skillsThe open agent skills tool - npx skills. Contribute to vercel-labs/skills development by creating an account on GitHub. 120
Joshua Iz @joshuaiz.com · 27/04/2026Using this patten I wrote about: joshuaiz.com/words/the-or... Along with a shared context doc in Linear per-project. It is actually working really well. Both Claudes update the shared doc so they are always in sync and don't lose context.joshuaiz.comJoshua Iz - Words - The Orchestrator/Builder PatternWriting, music, and ephemera from Joshua Iz 000
The Dark Prince of Donair @erisianrite.com · 27/04/2026MCP if I need distribution, oauth, tools, bells & whistles, etc., using github and the claude marketplace to manage skills for small groups, but no version pinning in projects is rough (maybe agents/claude.md would work here). 010
Bryan Russett @bryan.run · 27/04/2026all of the above to date, now working on generating dynamic skill files from agent traces plus external context 000
sabledima.bsky.social @sabledima.bsky.social · 28/05/2026the effect - when the cost of optimization increases an expected benefit from this optimization. many moments linked with llm optimisation & fomo around it remind me of it more often. e.g. "Claude, what is my schedule for today?" - "Master, you have a hair cut at 1pm only" 000
sabledima.bsky.social @sabledima.bsky.social · 28/05/2026everyone today is a overhyped with collecting & managing all these "SKILLs md", "build your 2nd brain", etc. Pursuing fomo we forget to live. And to think ourselves. back to the topic: I've downloaded /gstack by Garry Tan (YC) and tried it out of curiosity, but now magic happened expectedly :-)skill.mdAgent Skills Overview - Agent SkillsA standardized way to give AI agents new capabilities and expertise. 000
travis gerke @travisgerke.bsky.social · 27/04/2026I was struggling with this recently, had Claude walk me through a new strategy which (absurdly?) includes a skill for skills. Without getting into either of our specifics, a generic meta-skill pattern for this looks something like attached image. Interested in other replies here! 120
Sam Clemente @samclemente.me · 27/04/2026I’ve got a repo of project-agnostic skills that I’ve exposed via MCP using a tool I built called MyContextProtocol which helps with exposing the skills as reasonably usable MCP endpoints with metrics and all kinds of fun stuff The code is at github.com/Stygian-Tech... if you want to spin it upgithub.comGitHub - Stygian-Tech/my-context-protocolContribute to Stygian-Tech/my-context-protocol development by creating an account on GitHub. 100
holkerveen.bsky.social @holkerveen.bsky.social · 22/08/2026I use global skills and manage two sets: 1 for my professional setup, and 2 for my pet projects (different machine). That's enough management for me atm. I'd advise asking AI to update skills based on your sessions regularly. Makes it kind of self learning! Works ok for me. 010
Alex Marshall @alxm.me · 18/08/2026I use four layers: 1. Account-level CLAUDE.md for behavioural changes I always want from the LLM 2. Project-level CLAUDE.md 3. LLM-maintained documentation within every folder in the repo itself to document code and memorialise decisions 4. Skills only for narrow, repeated tasks, eg. reviews 100
Dario Djuric @dariodjuric.bsky.social · 28/04/2026I check the skills I’ve installed every once in a while and prune them. It’s difficult, as some are global and some are project-based. Would be great to have a UI to help manage them by highlighting the ones that are rarely used 000
tom larkworthy @larkworthy.bsky.social · 27/04/2026the main skill is /improve which writes and refactors the skills. I have knowledge directory for background context, and skills are for for small development workflows. These all live in a "developer" repo, and the real code repos are submoduled in, and worktree'd out. 220
dame @dame.is · 27/04/2026i don’t use skills, mcps, or anything, just raw doggin the LLM chat interface and avoiding new shiny objects like the plague 5380
Lawrence Jones @lawrencejones.dev · 27/04/2026We’re looking into this now for our AI SRE where we want to index people’s skills and either run them in an investigation ourselves or delegate to their infrastructure to run them. Sadly there is no real convention. Teams seem to do whatever works best for them, and can vary even in orgs. 100
sōm ⁂ @hyl.st · 27/04/2026dotfiles management for global skills, in repo ones are kept in sync via git :) 010
sɴᴅƨ @snds.design · 27/04/2026I've been using this in concert with Google Drive and Github. github.com/AgriciDaniel... It effectively turns your skills and session context into a super-wiki/Claude micro-brain.github.comGitHub - AgriciDaniel/claude-obsidian: Claude + Obsidian knowledge companion. Persistent, compounding wiki vault based on Karpathy's LLM Wiki pattern. /wiki /save /autoresearchClaude + Obsidian knowledge companion. Persistent, compounding wiki vault based on Karpathy's LLM Wiki pattern. /wiki /save /autoresearch - AgriciDaniel/claude-obsidian 040
Craig Hughes @craig.rungie.com · 27/04/2026“Global” and “project” are fine. My problem is with “iff typescript project” or “iff using excel” situations where I don’t want them globally or per project I want them by category. I (Claude) wrote a wrapper script for me that detects files like tsconfig or *.xls and loads in other directories. 210
Nick Gerakines @ngerakines.me · 27/04/2026I've been using chezmoi with a private git repository.chezmoi.iochezmoi - chezmoiManage your dotfiles across multiple machines, securely. 030
toni @tkukurin.bsky.social · 27/04/2026repo-specific: put superset of skills in a git repo, bootstrap per project by asking your coding agent to appropriately distribute (subset of) files. branch and upstream any updates. 110
Ilja @fubits.dev · 27/04/2026I took the first step to set up a repo for a particular set of skills (github.com/fubits1/svel...), but so far I failed to establish a practice of updating it. I probably need to block a fix time slot (maybe Friday mornings) in m calendar. 010
Jeremy Lewi @jeremy.lewi.us · 27/04/2026I think there is a fundamental UX issue with skills.md. skills.md are documentation for AIs and they have all the failure modes of documentation. The fundamental issue is that skills.md try to anticipate the future 'What will some unknown user need to know to accomplish some unknown task".skills.md 160
Jay Springett @thejaymo.net · 27/04/2026Sorry to send a 8k word long post. But I recently wrote this essay about lessons from interactive fiction engine design and agent harnesses, and give a concrete example of how to author better skill files thejaymo.net/2026/03/19/h...thejaymo.netHard Worlds For Little Guys - thejaymo.netWhat is it like to be an AI agent from the inside? This essay uses interactive fiction, MUDs, and world design to argue that agent environments are worlds, and that harder worlds make more reliable ag... 180
Shawn Simister @narphorium.com · 27/04/2026Separate repo for global skills. Only a few repo-specific skills. For atelier.dev, the customize skill keep hashes so it knows when the others need to be up updated 000
Boris @bmann.ca · 27/04/2026Check out @jeremie.com's vit v-it.org which is treating open source, especially today, as a network Skills in particular are a great use case for it.v-it.orgvit — open source is socialdevelopers and AI agents share software capabilities through a social network — discover what others have built, adapt it to your codebase, build reputation by vouching for what works. 070
