tune-workspace
Skill0 starsChanged 4 months agoMITView on GitHub
Turn a repeated correction or a proven workflow into the smallest written rule, in the right place, always shown to you first. For behavior rules, area/project shapes, and new skills — not for facts to remember (that's the remember skill).
In this repository17 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- add-areaSkill
- connect-toolsSkill
- new-projectSkill
- rememberSkill
- setup-workspaceSkill
- tune-workspaceSkill
- update-workspaceSkill
- workspace-helpSkill
- add-areaSkill
- connect-toolsSkill
- new-projectSkill
- rememberSkill
- setup-workspaceSkill
- update-workspaceSkill
- workspace-helpSkill
The file
We split this skill 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: line 20
Don't run it for a one-off preference — not everything needs to become a rule.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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
tune-workspace
This is how the workspace gets sharper over time without bloating into a pile of
rules — the rule-writing loop: turn a proven correction into the smallest written
rule, in the right place, and always show it before writing. (Facts, preferences, and
decisions about you or the business aren't rules — those are memory; the remember
skill keeps those.)
When to use
- The owner corrected the same thing twice.
- They said "do it this way from now on," "make that a rule," or "this keeps happening."
- A workflow worked well and they want it to be the default.
Don't run it for a one-off preference — not everything needs to become a rule.
How to run it
Name the lesson in one line. Propose the smallest written rule and where it belongs —
a durable behavior rule in AGENTS.md, the shape of an area or project in its
template, or a repeatable workflow as a new skill under skills/. Show the exact
change, wait for a yes, then write it. The two things that never bend: the smallest
change that captures the lesson, and never edit a core file without showing the change
first. If a lesson isn't stable enough to keep, say so and skip it.
One extra step when the rule becomes a new skill (a repeatable workflow saved
in its own folder under skills/): if this workspace ships hidden discovery mirrors,
copy the complete skill package into each mirror that exists —
.claude/skills/<name>/ (what Claude Code reads) and
.agents/skills/<name>/ (what Codex, Gemini CLI, Cursor, OpenCode, and
Devin/Windsurf read). Include SKILL.md and every disclosed reference, preserving
relative paths and byte identity with skills/<name>/. A partial package can point
at instructions the runtime cannot load. If a mirror root is absent, skip that one;
the canonical skill still runs when the owner asks in plain English.
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.

