extract-design-rationale
Skill1 starsChanged 4 months agoMITView on GitHub
Extract defensible design rationale from project notes, logs, specifications, research, and artifacts. Use when a case study shows what was built but not why, needs a constraint-alternatives-decision matrix, must explain rejected options and trade-offs, or needs to connect a UI, product, system, or AI-agent decision to evidence and measurable consequences.
In this repository11 other files it gives its agents.
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 13
Keep only decisions that changed user outcomes, system behavior, risk, cost, delivery, or learning.
- Distinctive: line 26
Do not invent alternatives to make a decision look rigorous.
≈ “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
Extract Design Rationale
Turn implementation history into a decision argument. Read references/rationale-matrix.md.
Workflow
- Inventory candidate decisions in the supplied evidence.
- Keep only decisions that changed user outcomes, system behavior, risk, cost, delivery, or learning.
- For each decision, extract:
- the constraint or problem;
- the tension between competing goals;
- the alternatives genuinely considered;
- why each alternative was rejected or deferred;
- the final decision and owner;
- the evidence available at decision time;
- the observed consequence or metric.
- Separate contemporary evidence from hindsight.
- Mark absent alternatives, rationale, ownership, or outcomes as evidence gaps.
- Rank the records by narrative importance.
Do not invent alternatives to make a decision look rigorous. Do not convert a team decision into sole-designer ownership. Do not claim business alignment without a named goal, constraint, or metric.
Output
Return:
- a one-sentence decision thesis;
- a ranked shortlist of pivotal decisions;
- one rationale matrix per decision;
- the strongest visual or artifact to pair with each record;
- unresolved evidence gaps.
Example
Input:
Users missed background-agent failures. We considered persistent alerts, a blocking
modal, and an activity timeline. Support logs showed most failures were recoverable.
We shipped the timeline plus a thresholded alert; failed-task recovery rose from 62%
to 84% in the beta.Expected outcome: a matrix naming failure visibility as the constraint, the modal and persistent-alert trade-offs, the selected timeline-plus-threshold model, the support-log evidence, and the 62%→84% recovery result.
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.

