flutter-capture-pattern
Skill2 starsChanged 41 days agoMITView on GitHub
Record a Flutter convention or pattern proven in project code as a reference in the flutter-llm-toolkit. Use when asked to capture, document or extract a pattern into the toolkit.
In this repository4 other files it gives its agents.
- AGENTS.mdAGENTS.md
- flutter-code-qualitySkill
- flutter-conventionsSkill
- flutter-testingSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 2 are distinctive, and 1 is 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.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 12
Write to a checkout of the toolkit repository, not to a project's installed copy.
- Distinctive: line 25
No project names, repository paths or developer-specific setup appear.
- Distinctive: line 26
Every rule appears once; no absolutes except real invariants.
≈ “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
Capture a Flutter Pattern
Turn something proven in a project into toolkit knowledge: a new reference file in flutter-conventions/references/, an addition to an existing one, or a new rule in the flutter-code-quality checklist. The toolkit is the single source; project copies are refreshed from it with the README update prompt.
Where the toolkit lives
Write to a checkout of the toolkit repository, not to a project's installed copy. The README install prompt clones to ~/toolkits/flutter-llm-toolkit; if no checkout is known, ask where it is. Read the toolkit's AGENTS.md for the writing rules before drafting.
What qualifies
A pattern earns a place when it recurs across features or apps, a model would not produce it unguided, and the source code demonstrates it working. One-off product logic, unfinished code and project-specific configuration do not. If the candidate is a single lint-like rule with a clear bad and good form, it belongs in the checklist under the matching section rather than in a reference.
Working from source
Read the implementation the pattern comes from, not a description of it. Generalise names, trim product fields and logging, keep real signatures of shared helpers, and name files by their place in the layout (lib/common/util/debouncer.dart) rather than by the project they came from. Where the code deviates from its own convention, document the convention and say nothing about the deviation. When the pattern overlaps an existing reference, extend that file instead of creating a neighbour, and update the "Where to Look" table in flutter-conventions/SKILL.md only when a new file was added.
Done when
- The reference follows the format of its neighbours (H1, one-line scope, H2 sections, code carries the content, optional closing Anti-Patterns).
- No project names, repository paths or developer-specific setup appear; the toolkit is used by people without access to the source project.
- Every rule appears once; no absolutes except real invariants; no process scripts.
- A reader with the file alone could reproduce the pattern in a new app.
- You report the file path, what was added, and whether the project copy needs refreshing.
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.

