gaboniogabonio/agent-operator-toolkit/AGENTS.md

agent-operator-toolkit

AGENTS.md1 starsChanged 3 months agoMITView on GitHub

These instructions apply anywhere this toolkit is checked out. AGENTS.md is the canonical instruction source for this repo. The CLAUDE.md, CODEX.md, and GEMINI.md compatibility files point back here instead of containing divergent rules. This repo is a portable multi-agent marketplace for reusable AI-development plugins and skills. Keep distribution infrastructure generic, keep each plugin self-contained under plugins/, and use native Codex, Claude Code, GitHub Copilot, and Gemini CLI surfaces. Project-specific requirements, credentials, proprietary systems, employer names, and business rules belong in…

Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,774 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository11 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- adversarial-reviewSkill
- agent-operator-claude-setupSkill
- agent-operator-codex-setupSkill
- agent-operator-gemini-setupSkill
- agent-operator-setupSkill
- builder-kitSkill
- develop-from-specSkill
- marketplace-managementSkill
- pr-review-loopSkill
- start-new-appSkill

The file

Show its rules and text

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

13 rules. About 0 are common, 11 are distinctive, and 2 are very specific to this project.

≈0
Common · estimate
Many files like this have a rule of this kind.
≈11
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.

Common
Distinctive
Very specific

Distinctive:
Before changing this repo:
line 15
Distinctive:
read WORKFLOWS.md for plugin or release work
line 18
Very specific:
read the target plugin's SKILL.md and directly referenced guidance
line 19
Distinctive:
do not make one plugin depend on an organization-only bot, slash command, or service
line 45
Distinctive:
Use the docs like this:
line 49
Distinctive:
When new knowledge is learned well enough to keep:
line 61
Distinctive:
add general marketplace policy to this file
line 63
Distinctive:
add repeated marketplace maintenance patterns to WORKFLOWS.md
line 64
Distinctive:
add setup or usage changes to README.md
line 65
Very specific:
keep README-CODEX.md, README-CLAUDE.md, README-COPILOT.md, and README-GEMINI.md aligned with native client commands
line 66
Distinctive:
keep plugin behavior in the owning plugin
line 67
Distinctive:
keep project-specific rules in the project that owns them
line 68
Distinctive:
Do not duplicate the same guidance across many repos when it can live once here.
line 70

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 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

Agent Operator Marketplace AGENTS

These instructions apply anywhere this toolkit is checked out.

AGENTS.md is the canonical instruction source for this repo. The CLAUDE.md, CODEX.md, and GEMINI.md compatibility files point back here instead of containing divergent rules.

Purpose

This repo is a portable multi-agent marketplace for reusable AI-development plugins and skills. Keep distribution infrastructure generic, keep each plugin self-contained under plugins/, and use native Codex, Claude Code, GitHub Copilot, and Gemini CLI surfaces.

Project-specific requirements, credentials, proprietary systems, employer names, and business rules belong in the project or private plugin that owns them, not here.

Read Order

Before changing this repo:

read README.md
read WORKFLOWS.md for plugin or release work
read the target plugin's SKILL.md and directly referenced guidance

Scope

Track changes here only for:

reusable, organization-neutral plugins under plugins/
Codex and shared Claude Code/GitHub Copilot marketplace registries
Gemini CLI extension metadata, commands, and skill entrypoints
marketplace contributor documentation
lightweight validation and release automation

Do not track:

project-specific business logic
vendored application or business-logic source unrelated to a marketplace plugin
secrets or credentials
machine-specific operational notes
internal commands, private workflows, or proprietary platform guidance
generated reports or scratch outputs

Plugin boundaries

the target project owns product requirements and project-specific instructions
each plugin owns only its reusable workflow and references
marketplace root files own discovery, validation, and publishing behavior
do not make one plugin depend on an organization-only bot, slash command, or service

Documentation Governance

Use the docs like this:

README.md: marketplace usage and structure
WORKFLOWS.md: plugin addition and release workflows
.claude-plugin/marketplace.json: shared Claude Code and GitHub Copilot marketplace registry
.agents/plugins/marketplace.json: Codex marketplace registry
gemini-extension.json, commands/, and skills/: Gemini CLI extension surface
plugins/: installable plugin roots
scripts/check: marketplace validation

Documentation Maintenance Rules

When new knowledge is learned well enough to keep:

add general marketplace policy to this file
add repeated marketplace maintenance patterns to WORKFLOWS.md
add setup or usage changes to README.md
keep README-CODEX.md, README-CLAUDE.md, README-COPILOT.md, and README-GEMINI.md aligned with native client commands
keep plugin behavior in the owning plugin
keep project-specific rules in the project that owns them

Do not duplicate the same guidance across many repos when it can live once here.

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.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

It worked
Partly
It didn't

What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

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.

Check your setup, free