agentleFS
eduardo-sleduardo-sl/go-agent-skills/AGENTS.md

go-agent-skills

AGENTS.md85 starsChanged 18 days agoMITView on GitHub

You are working on go-agent-skills, a repository of curated AI agent skills for Go development. These skills follow the Agent Skills standard format. This is NOT a Go application. It is a collection of Markdown instruction files (SKILL.md) that teach AI coding agents how to write idiomatic Go. The repo has no Go source code, no go.mod, and no build artifacts. 1. **Read docs/SKILL_GUIDELINES.md first.* It defines the quality standards. 2. *Keep SKILL.md within the 250-line budget* (CI warns above…

We split this AGENTS.md into the rules it gives an agent

6 rules. About 0 are common, 2 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    YAML frontmatter is mandatory: Every SKILL.md must have name, description, user-invocable, license, compatibility, allowed-tools, …

    line 28
  2. Very specific:

    Create skill-name/SKILL.md with proper frontmatter.

    line 47
  3. Distinctive:

    Add the skill to the README.md catalog table.

    line 48
  4. Very specific:

    If creating a new category, add _category.json.

    line 49
  5. Very specific:

    Run ./scripts/validate.sh to verify format.

    line 50
  6. Distinctive:

    No trailing whitespace. Final newline on all files.

    line 57

≈ “Common” is an estimate: at least 1 in 5 of the 36,368 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
AGENTS.md

Guidance for AI Coding Agents

You are working on go-agent-skills, a repository of curated AI agent skills for Go development. These skills follow the Agent Skills standard format.

Project Context

This is NOT a Go application. It is a collection of Markdown instruction files (SKILL.md) that teach AI coding agents how to write idiomatic Go. The repo has no Go source code, no go.mod, and no build artifacts.

Repository Structure

skills/(category-name)/skill-name/SKILL.md
  • Categories use parentheses: (code-quality), (architecture), etc.
  • Each category has a _category.json with metadata.
  • Each skill is a directory containing at minimum a SKILL.md.
  • Skills may also contain references/ for supplementary docs.

When modifying skills

  1. Read docs/SKILL_GUIDELINES.md first. It defines the quality standards.
  2. Keep SKILL.md within the 250-line budget (CI warns above 250 and fails above 500). Move detailed content to references/.
  3. YAML frontmatter is mandatory. Every SKILL.md must have name, description, user-invocable, license, compatibility, allowed-tools, metadata.author and metadata.version. allowed-tools is least-privilege: no Edit/Write for review-only skills, and every Bash(...) entry scoped to one binary.
  4. Name must match directory name. go-code-review/SKILL.md → name: go-code-review.
  5. Description must include trigger phrases AND negative triggers. Example:
    Use when: "review this code", "check this PR"
    Do NOT use for: security audits (use go-security-audit)
  6. Run validation before committing:
    ./scripts/validate.sh

When creating new skills

  1. Choose the correct category directory or create a new one.
  2. Create skill-name/SKILL.md with proper frontmatter.
  3. Add the skill to the README.md catalog table.
  4. If creating a new category, add _category.json.
  5. Run ./scripts/validate.sh to verify format.

Code style for this repo

  • Markdown: ATX-style headers (#), fenced code blocks with language tags.
  • Shell scripts: set -euo pipefail, shellcheck-clean.
  • JSON: 2-space indentation.
  • No trailing whitespace. Final newline on all files.

Commit conventions

Follow Conventional Commits:

feat(skill): add go-grpc-patterns skill
fix(go-error-handling): correct errors.Join example for Go 1.20+
docs(readme): add installation instructions for Windsurf
chore(ci): add SKILL.md line count validation

Distribution

This repo is distributed via the npx skills CLI. Users install with:

npx skills add eduardo-sl/go-agent-skills

The CLI discovers skills by recursively searching skills/ for SKILL.md files.

Platform discovery files

Each agent has its own discovery mechanism. When adding or removing a skill, update ALL of these:

FilePlatformPurpose
AGENTS.mdUniversal (Codex, Gemini CLI, Factory)Agent instructions for this repo
CLAUDE.mdClaude CodeSession context with skill list
.claude-plugin/marketplace.jsonClaude Code marketplacePlugin manifest with skill paths
.cursor/rules/go-skills.mdcCursorProject rules with skill discovery
.windsurf/rules/go-skills.mdWindsurfCascade rules
.clinerulesCline / Roo CodeProject rules
.github/copilot-instructions.mdGitHub CopilotCustom instructions
.opencode/config.jsonOpenCodeConfig manifest

The skill list in all files MUST match. If a file lists a skill that doesn't exist (or omits one that does), users on that platform get confused.

Language

  • All files in this repository must be written in English. This includes SKILL.md files, documentation, commit messages, comments, and any other text content. No exceptions.

Important constraints

  • Skills must be agent-agnostic. No Claude-specific, Cursor-specific, or Copilot-specific instructions inside SKILL.md files. Skills work everywhere.
  • Skills must be self-contained. No skill should reference or depend on another skill being loaded. Each skill stands alone. Naming another skill as a pointer is fine — negative triggers already do it — but no skill may require another one to have been loaded first. go-skills-router is the one index skill: its whole job is naming the others. It still stands alone, and nothing depends on it.
  • Do not add Go source code to this repo. Code examples belong inside SKILL.md files as fenced blocks, not as separate .go files.
  • Do not add package managers, build tools, or frameworks. This repo is distributed by copy/symlink. Keep it simple.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

Reviews can't be read right now.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.