lyoneellyoneel/ly-agent-skills/AGENTS.md

ly-agent-skills

AGENTS.md0 starsChanged 2 months agoMITView on GitHub

Repository for AI agent skill definitions. The content is Markdown-based instruction files consumed by Crush at runtime — there is no build system, no source code, and no test suite. Markdown-only repository. Each top-level subdirectory is one skill. Skills are loaded by name into Crush's context window as system prompt extensions. The skill inventory lives in the README.md Skills section; keep it current when adding or removing a skill. The Skills section consists of a summary table followed by one…

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 repository9 other files it gives its agents.
- agentSkill
- agents-md-syncSkill
- aur-pkg-analysisSkill
- commitSkill
- crush-sessionSkill
- gen-agentSkill
- git-aware-mvSkill
- tg-notifySkill
- tg-notipySkill

The file

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

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.

Common
Distinctive
Very specific

Distinctive:
After the overview paragraph, add two links on one line:
line 52

≈ “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

Repository for AI agent skill definitions. The content is Markdown-based instruction files consumed by Crush at runtime — there is no build system, no source code, and no test suite.

Project Type

Markdown-only repository. Each top-level subdirectory is one skill. Skills are loaded by name into Crush's context window as system prompt extensions.

Repository Structure

skill-name/
  SKILL.md       # YAML frontmatter + step-by-step instructions (the skill)
  README.md      # public-facing summary (consumed by humans / mirrors)
  references/    # detailed guides and workflow docs (loaded progressively)
  templates/     # copyable skeletons and boilerplate
  scripts/       # executable automation (bash, python, etc.)
  assets/        # static data, images, lookup tables

The skill inventory lives in the README.md Skills section; keep it current when adding or removing a skill.

README Skills Section Format

The Skills section consists of a summary table followed by one overview section per skill.

Summary table (directly under ## Skills):

Three columns: Skill, Description, README.
Skill column: the folder name, linked to the skill's detail anchor ([<folder-name>](#<display-name>---<folder-name>)).
Description column: a one-line, very brief statement of what the skill does; a short fragment, not the overview paragraph. It may carry inline links to a skill's upstream repo. Keep it terse.
Where a skill's overview paragraph is a single terse sentence, that same sentence is reused verbatim as the description; otherwise the description is distilled from the capability list, never copied from a longer overview paragraph.
README column: a single link whose visible text is DOC, wrapped in literal square brackets with a space inside each bracket, pointing to the skill's README: [ [DOC](<folder-name>/README.md) ].

Detail sections (one per skill, after the table):

Heading is the display name followed by the folder name in brackets: ### <Display Name> - [<folder-name>].
Body is the skill's overview paragraph, copied verbatim from the opening paragraph of its OWN README.md (the single source of truth). It is not a new summary and not the skill's full README content.
After the overview paragraph, add two links on one line:
- an up link back to the summary table, with visible brackets and spaces: [ [UP to skills table](#skills-table) ];
- a more-info link to the skill's own README, with visible brackets and spaces: [ [More info in README](<folder-name>/README.md) ]. The table carries an anchor #skills-table placed directly under the ## Skills heading.

Key Conventions

Paths always relative.
Skill format and contribution rules: see CONTRIBUTING.md (frontmatter shape, sub-5000-token SKILL.md body, no markup in skill markdown, Python for scripts, Conventional Commits).

Git & Branching

Default branch: master

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.

Check your setup, free