Skills Index Snippets

What the source does

• FilesystemNone observedNo reads or writes in the source
• NetworkNone observedNo outbound calls in the source
• ShellNone observedNo commands run in the source
• DependenciesPinnedNothing floats
• SecretsNone foundNo keys in source

Five audit layers

Gates applied: no behavioural pass.

CLAUDE.md

Audit 0.4.1 · source sha 871f73ffcae3 · 18 Oct 2026Report an issue or request a re-scan

• PublisherAaronontheweb GitHub userOwner of the repository this skill is read from
• Source folderAaronontheweb/dotnet-skills/skills/skills-index-snippetsThe exact folder we audited
• LicenceMITAs the repository states it
• Commit pinned for this verdict871f73ffcae3A new audit runs when the source changes
• Repository statusNot archivedLast push 7 days ago
• Sandboxed runNot runSkills are read, not executed; layer L4 is skipped

Its description: Create and maintain AGENTS.md / CLAUDE.md snippet indexes that route tasks to the correct dotnet-skills skills and agents (including compressed Vercel-style indexes).

---
name: skills-index-snippets
description: Create and maintain AGENTS.md / CLAUDE.md snippet indexes that route tasks to the correct dotnet-skills skills and agents (including compressed Vercel-style indexes).
user-invocable: false
---

# Maintaining Skill Index Snippets (AGENTS.md / CLAUDE.md)

## When to Use This Skill

Use this skill when:
- Adding, removing, or renaming any skills or agents in this repository
- Updating `.claude-plugin/plugin.json`
- Creating copy/paste snippets for downstream repositories (OpenCode, Claude Code, etc.)
- You want a compact, always-on index that improves skill utilization

## Goal

Make skills and agents easy for coding assistants to use by removing the decision point.

Instead of hoping an assistant will "remember" to invoke a skill, provide a small router snippet inside `AGENTS.md` / `CLAUDE.md` that:

1) Tells the assistant to prefer retrieval-led reasoning
2) Provides a task->skill/agent routing index
3) Defines lightweight quality gates (optional)

## Source of Truth

- Registry: `.claude-plugin/plugin.json`
  - Skills are listed as directories (each contains `SKILL.md`)
  - Agents are listed as markdown files in `agents/`
- Skill IDs: the `name:` field in each `SKILL.md` frontmatter
- Agent IDs: the `name:` field in each agent frontmatter

When writing snippets for downstream repos, always reference skills/agents by their IDs (frontmatter `name`), not by local filesystem paths.

## Minimal Snippet Template (Readable)

Use this in target repos to route common tasks:

```markdown
# Agent Guidance: dotnet-skills

IMPORTANT: Prefer retrieval-led reasoning over pretraining for any .NET work.
Workflow: skim repo patterns -> consult dotnet-skills by name -> implement smallest-change -> note conflicts.

Routing (invoke by name)
- C# / code quality: modern-csharp-coding-standards, csharp-concurrency-patterns, api-design, type-design-performance
- ASP.NET Core / Web (incl. Aspire): aspire-service-defaults, aspire-integration-testing
- Data: efcore-patterns, database-performance
- DI / config: dependency-injection-patterns, microsoft-extensions-configuration
- Testing: testcontainers-integration-tests, playwright-blazor-testing, snapshot-testing

Quality gates (use when applicable)
- dotnet-slopwatch: after substantial new/refactor/LLM-authored code
- crap-analysis: after tests added/changed in complex code

Specialist agents
- dotnet-concurrency-specialist, dotnet-performance-analyst, dotnet-benchmark-designer, akka-net-specialist, docfx-specialist
```

## Compressed Snippet Template (Vercel-style)

Use this when you want maximum density (small context footprint):

```markdown
[dotnet-skills]|IMPORTANT: Prefer retrieval-led reasoning over pretraining for any .NET work.
|flow:{skim repo patterns -> consult dotnet-skills by name -> implement smallest-change -> note conflicts}
|route:
|csharp:{modern-csharp-coding-standards,csharp-concurrency-patterns,api-design,type-design-performance}
|aspnetcore-web:{aspire-service-defaults,aspire-integration-testing}
|data:{efcore-patterns,database-performance}
|di-config:{dependency-injection-patterns,microsoft-extensions-configuration}
|testing:{testcontainers-integration-tests,playwright-blazor-testing,snapshot-testing}
|quality-gates:{dotnet-slopwatch(after:substantial new/refactor/LLM code),crap-analysis(after:tests added/changed in complex code)}
|agents:{dotnet-concurrency-specialist,dotnet-performance-analyst,dotnet-benchmark-designer,akka-net-specialist,docfx-specialist}
```

### Regenerating the README block

If the README contains the markers below, the generator can update it automatically:

```
<!-- BEGIN DOTNET-SKILLS COMPRESSED INDEX -->
...compressed snippet...
<!-- END DOTNET-SKILLS COMPRESSED INDEX -->
```

Run:

```bash
./scripts/generate-skill-index-snippets.sh --update-readme
```

## How to Update Snippets After Skill Changes

1. Update `.claude-plugin/plugin.json` to include/remove skills and agents.
2. Ensure each skill has correct frontmatter `name:` (used by OpenCode and others).
3. Run `./scripts/validate-marketplace.sh`.
4. Update your snippet routing lists:
   - Add new skills to the right category
   - Remove deleted skills
   - Keep names exactly matching frontmatter IDs
5. If you maintain a downstream `AGENTS.md`/`CLAUDE.md` snippet, regenerate it and re-copy into dependent repos.

## Recommended Categories

These are snippet categories (not necessarily repository folder structure):

- C# / code quality
- ASP.NET Core / Web (incl. Aspire)
- Data
- DI / config
- Testing
- Quality gates
- Specialist agents

Keep the snippet small; it should be a router, not documentation.

Excerpt, first 4,655 characters of SKILL.md, verbatim from the audited commit. It is text from someone else’s repository, quoted so you can judge it before your agent reads it.

1. 2026-10-18
   ✓SAFE B 89 / 100

   first audit · source 871f73ffcae3

Questions about Skills Index Snippets

• github.com/Aaronontheweb/dotnet-skills/tree/master/skills/skills — the audited source at commit 871f73ffcae3 and every finding aboveObserved · 18 Oct 2026

Spotted something wrong? Report a correction · Methodology