compare-states
Skill3 starsChanged 37 days agoMITView on GitHub
Build one self-contained HTML page that sets two or more states of the same mechanism side by side at full identifier-level detail.
In this repository5 other files it gives its agents.
- AGENTS.mdAGENTS.md
- explain-simplySkill
- full-pictureSkill
- greenfieldSkill
- options-analysisSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 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.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 23
Read the code that produces each box before drawing it.
- Distinctive: line 28
Never drop a box to make it fit.
≈ “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
Produce ONE self-contained HTML file (Mermaid via CDN) and deliver it rendered in the side panel
(SendUserFile with display: render; where that tool is absent, name the path). Put it beside the
session's other docs written for the user.
Three regions, left to right:
- Shared — spans every row, split into two panels of its own. Left: the problem in three sentences, then a table of every object the diagrams name. Right: a glossary of every label, grouped tables · functions · columns · arguments, one plain sentence each.
- Diagrams — one Mermaid diagram per state, stacked. The same boxes and the same layer count in every one, so they overlay.
- Reading — one panel per diagram, in the same grid row so the tops align: the state's name, a numbered list that walks the arrows, and one coloured verdict box.
Rules:
- Read the code that produces each box before drawing it. A doc's description is not a source. Every line number comes from a file opened this session.
- Real identifiers inside a diagram — table, id argument, columns. Plain words live in the glossary and the reading. Never paraphrase inside a box.
useMaxWidth: false, so every diagram renders at one scale. When one overflows, widen the page. Never drop a box to make it fit.- A numbered arrow is one statement the code runs, in order. An unnumbered arrow is a pointer that exists once the statements have run. Colour only the row under discussion.
- The chat reply is a four-line reading order plus the next action. No ASCII diagram in chat, no
.mdfallback.
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.

