Skip to main content

A weekly stocktake for your skills

After my posts about automations, I got some questions about other examples where I use this feature. One I like to share is a weekly check that validates my current list of installed skills.

Why do we need this?

Skills are easy to add and easy to forget. Over time you collect skills that overlap, skills that point to tools that changed, and skills you no longer use.

I wanted a recurring check that tells me how to cleanup my skills list.

Starting from an existing skill

I didn't start from scratch. The ECC repository contains a skill-stocktake skill for Claude Code. It's a /skill-stocktake slash command that audits the skills in ~/.claude/skills/ and the project-level .claude/skills/, and it has two modes:

  • Quick Scan: re-evaluates only the skills that changed since the last run.
  • Full Stocktake: a complete review.

Under the hood it uses shell scripts, a results.json cache, and subagents that evaluate the skills in batches of about 20.

That's a good design for Claude Code. As I use GitHub Copilot most of the time at work, I ported and adapted it to work for GitHub Copilot as well: skill-stocktake in wullemsb/skills.

What I changed

  • Scope: the port scans skills/*/SKILL.md, plugins/*/skills/*/SKILL.md and .github/copilot/skills/*/SKILL.md, plus any paths I provide. It reports which of these locations it found.
  • No scripts, cache or subagents: the skill is plain instructions. That also means no Quick Scan, so every run is a full pass. I’ll probably re-introduce the scripts and the usage of subagents in a later iteration.
  • Usage signals instead of 7d/30d counts: the original knows how often a skill was used. The port looks for evidence in the repository (documentation, examples, tests, settings). If it can't find reliable evidence, it says Unknown instead of guessing.
  • Overlap checks: it compares skills with each other and with README.md, .github/copilot/, AGENTS.md, CLAUDE.md and MEMORY.md, if present.
  • Freshness checks: it verifies technical references with whatever web lookup the current client offers.

What I kept

The four phases stay the same: inventory, quality evaluation, summary table and consolidation. So do the verdicts:

Verdict Meaning
Keep Useful and current
Improve Worth keeping, but specific improvements needed
Update Referenced technology is outdated
Retire Low quality, stale, misleading, or cost-asymmetric
Merge into [X] Substantial overlap with another skill; name the merge target

What I like most is that every verdict needs a self-contained reason. "Unchanged" is not allowed. For a Retire it has to name the defect and what already covers the need. For a Merge it has to name the target and what content should move. That's the difference between a report you can act on and a list of labels.

Remark: the skill never deletes, merges or rewrites skill files unless I explicitly ask for it. That's what makes it safe to run unattended.

Running it every week

I run this skill every week as a GitHub Copilot App automation. If you want the exact steps how to set this up, check out my previous post about Github Copilot App Automations.

Reading the result

You get a summary table with one row per skill:

 


That's it! The automation does the boring inventory. The decision to retire or merge stays with me.

More information

Labels:

Popular posts from this blog

Podman– Command execution failed with exit code 125

After updating WSL on one of the developer machines, Podman failed to work. When we took a look through Podman Desktop, we noticed that Podman had stopped running and returned the following error message: Error: Command execution failed with exit code 125 Here are the steps we tried to fix the issue: We started by running podman info to get some extra details on what could be wrong: >podman info OS: windows/amd64 provider: wsl version: 5.3.1 Cannot connect to Podman. Please verify your connection to the Linux system using `podman system connection list`, or try `podman machine init` and `podman machine start` to manage a new Linux VM Error: unable to connect to Podman socket: failed to connect: dial tcp 127.0.0.1:2655: connectex: No connection could be made because the target machine actively refused it. That makes sense as the podman VM was not running. Let’s check the VM: >podman machine list NAME         ...
Read more

Cache stampede: when our cache turned against us

While investigating some performance issues, we ran into an ASP.NET Core API that cached a fairly expensive aggregation query for 60 seconds. Under normal load, that was fine: one request rebuilds the cache, everyone else reads from it. Under peak load, dozens of requests would arrive in that same expiry window, all see a cache miss, and all fire the same expensive query in parallel. The database didn't like that. That was the moment when our caching layer stopped helping and started hurting. A burst of requests comes in at the same time, all miss the cache, and all go hammer the database or the downstream API at once. That's a cache stampede . The cache was supposed to protect our backend, and for a few hundred milliseconds it did the opposite. Why this happens IMemoryCache.GetOrCreate (and its async sibling) looks like it protects you, but it doesn't add any locking on its own. Look at the naive version: public async Task<Report> GetReportAsync(string key) ...
Read more

The role of ActivitySource in OpenTelemetry for .NET

While doing some pair programming to integrate OpenTelemetry tracing to a .NET application, we had a discussion on how to use the ActivitySource . It looks simple. You new one up, give it a name, start an activity, done. The discussion started when we added a second ActivitySource with the exact same name in a different class. This made us wonder: "Are we duplicating traces now? Is this a memory leak? Do we need a singleton?" So we decided to dig deeper. This post is what we learned… What ActivitySource actually is ActivitySource is part of System.Diagnostics , not part of the OpenTelemetry NuGet packages. Microsoft built tracing primitives directly into the BCL, and OpenTelemetry's .NET SDK simply listens to them. This is why you can add distributed tracing to a library without taking a dependency on OpenTelemetry at all. An ActivitySource is a factory for Activity objects, and an Activity is .NET's name for what OpenTelemetry calls a span.(don’t ask m...
Read more