skill-dict
Skill1 starsChanged 2 months agoMITView on GitHub
Use when the user types `/skill-dict`, asks about their personal skill library at ~/skills-library/, asks "what does <skill> do" about a library entry, asks to sync the library after installing a plugin, or asks to add a skill to the library.
In this repository5 other files it gives its agents.
- grill-meSkill
- newprojectSkill
- pickupSkill
- putdownSkill
- takenotesSkill
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 1 is distinctive, and 4 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.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 40
If multiple matches, ask via AskUserQuestion.
- Very specific: line 41
If no match, tell the user and suggest /skill-dict list to see what's available.
- Very specific: line 42
Read just that one file with Read and print it.
- Distinctive: line 48
Read-only audit comparing installed plugin versions to upstream sources.
- Very specific: line 59
Preserve hand-edits: When sync finds an existing file, do not modify it — even if the manifest disagrees on installed: or version:.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill 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.
Files
/skill-dict — Personal skill library at ~/skills-library/
You maintain a hand-curated catalog of every Claude Code skill you've installed or authored. One markdown file per parent unit (a plugin or top-level user-authored skill). Sub-skills inside a plugin live inside the parent's entry, not as separate files.
Library root: ~/skills-library/
skills-library/
├── README.md ← overview, conventions, entry template
├── INDEX.md ← scannable table of all entries
├── plugins/ ← installed from marketplaces or npm
│ └── *.md
└── authored/ ← your own skills
└── *.mdThe skill has 5 subcommands. Pick the one matching the user's argument; if none was given, ask which.
Subcommands
list (inline)
Print the contents of ~/skills-library/INDEX.md and stop. Do NOT read other files. The index is a single markdown table — that's the whole point of having it.
show <name> (inline)
- Locate the file: try
plugins/<name>.mdfirst, thenauthored/<name>.md. UseGlobif the user used a partial name. - If multiple matches, ask via
AskUserQuestion. - If no match, tell the user and suggest
/skill-dict listto see what's available. - Read just that one file with
Readand print it. Do NOT read other library files.
sync
Reconcile the library against installed plugins + user-authored skills. Idempotent. See references/sync.md for the full algorithm.
check-updates
Read-only audit comparing installed plugin versions to upstream sources. See references/check-updates.md for the full flow.
add <name>
Interactive guided add for self-authored skills. See references/add.md for the questions and template.
General rules
- Never read more than one library entry file in a single invocation unless the user explicitly asks for "all entries" — the whole point is to keep context cost low.
- Never delete or move existing library files without explicit user confirmation.
synconly adds. - Never auto-fill the user's "Where I use it" / "Why I kept it" sections from external assumption — those are your own notes. Stubs leave them empty.
- No git commits. The CLAUDE workspace root is not a git repo. Don't try.
- Preserve hand-edits. When
syncfinds an existing file, do not modify it — even if the manifest disagrees oninstalled:orversion:. your edits win.
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
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.

