dx-lesson
Skill8 starsChanged 4 months agoMITView on GitHub
Record one finding — a warning or a decision-with-rationale — as an append-only entry in foundation/lessons.md.
In this repository21 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- dx-archiveSkill
- dx-brainstormSkill
- dx-diagnoseSkill
- dx-distillSkill
- dx-domain-discoverSkill
- dx-domainSkill
- dx-frameSkill
- dx-implementSkill
- dx-initSkill
- dx-newSkill
- dx-planSkill
- dx-refactor-discoverSkill
- dx-referencesSkill
- dx-researchSkill
- dx-reviewSkill
- dx-review-triageSkill
- dx-roadmapSkill
- dx-standards-discoverSkill
- dx-standards-updateSkill
- dx-tddSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
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
- Distinctive: line 19
Use it for the lesson entry shape and to judge whether this is a lesson at all (a one-off finding is fine here.
- Distinctive: line 30
Keep it tight — the title is what future skills scan first.
- Very specific: line 34
If the finding looks like it will keep coming up across changes, note it — that is the trigger for graduating it into a standard via …
- Distinctive: line 45
Stop. Do not chain into another skill.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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
dx-lesson
Capture one finding so a future change re-reads it instead of relitigating it. A lesson is scar tissue or a load-bearing decision — a rule, not a story. Two flavors, both belong here: cautionary ("this broke because…") and decisional ("we chose X over Y because Z", including a rejected refactor: "don't re-deepen X — it's shallow on purpose because Y"). This is the anteroom to a standard, not an ADR — the framework has no separate decision register.
Guard. foundation/lessons.md must exist (it does after /dx-init). If it is missing, tell the user to run /dx-init first. Read foundation/glossary.md for naming — a one-line habit, no section.
1 — Get the finding
Take the finding from the argument. If nothing was passed, ask what broke or what was decided, and why it matters — one question, then proceed. Don't turn a capture into an interview.
2 — Load the entry shape (invoke dx-references with knowledge-layer)
Use it for the lesson entry shape and to judge whether this is a lesson at all (a one-off finding is fine here; a recurring one is a promotion candidate — see step 4).
3 — Append (never rewrite)
Append to the end of foundation/lessons.md. Never edit, reorder, dedupe, or reformat existing entries — the file is append-only; that friction is the point. One entry, this shape:
## <short title> — <YYYY-MM-DD>
<what happened / what was decided>. **Why:** <the reason it matters>. **How to apply:** <the rule going forward>.Keep it tight — the title is what future skills scan first.
4 — Flag for promotion if it will recur
If the finding looks like it will keep coming up across changes, note it — that is the trigger for graduating it into a standard via /dx-standards-update. Don't promote here; lessons are the append-only home.
Done when
The entry is the last section of foundation/lessons.md and existing entries are untouched. Then print and stop:
Lesson recorded: foundation/lessons.md — ## <short title>
Next: /dx-standards-update — only if this will recur across changes (promote it to a standard)Stop. Do not chain into another skill. One capture per invocation — batching invites half-written entries.
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.

