normative-design
Skill8 starsChanged 57 days agoMITView on GitHub
Legitimacy is a design property, not an afterthought. Use when a rule, policy, or allocation must be written or applied and the open question is "is this fair and consistently applied?" rather than "does it work", when a general rule does not determine a specific case, when harm emerged with nobody having decided it, or when "we've always done it this way" is the only justification on offer.
In this repository16 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- boundary-designSkill
- causal-auditSkill
- decision-bias-checkSkill
- estimationSkill
- evidence-synthesisSkill
- experience-and-transmissionSkill
- failure-forensicsSkill
- formal-correctnessSkill
- measurement-disciplineSkill
- narrative-sensemakingSkill
- problem-reframingSkill
- representation-and-possibilitySkill
- structure-discoverySkill
- systems-leverageSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
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 42
If no shape above matches, do not force a genius — use a built-in agent with the closest GENIUS: pattern (rules/agent-routes.json, INDEX.md …
≈ “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
Normative Design
Problem shape: a rule, policy, default, or allocation is being designed, applied, or defended, and the load-bearing question is normative rather than technical — who bears the cost, whether the rule determines this case at all, whether responsibility dissolved across a chain of individually-reasonable steps, and whether the arrangement is actually justified or merely inherited.
Relevant geniuses
| Agent | Use when |
| rawls | legitimate values collide (privacy vs security, fairness vs efficiency) — design behind the veil of ignorance, without knowing which party you will be; check the arrangement against the worst-off position |
| hart | a general rule does not clearly determine the specific case — analyze the open texture, reason by precedent, separate the ratio decidendi from what merely accompanied it |
| arendt | systemic harm occurred and no one intended it — audit for suppressed judgment and cog-in-machine structure before attributing malice |
| foucault | the only answer to "why this way?" is "we've always done it this way" — trace the contingent origin of what presents itself as natural; map who the practice makes into what |
Invocation
- Pick the best-fit agent above. If two or more fit, run
tools/genius-invoker.sh route "<problem>"and take the top ranked match. - Load it:
tools/genius-invoker.sh invoke <agent> "<problem>", then readagents/genius/<agent>.mdin full. - Apply the agent's
<workflow>step by step — do not skip steps — and answer in its<output-format>. The deliverable names who bears which cost under the proposed rule, not just that the rule is defensible. - Typical chain: foucault denaturalizes the inherited practice → rawls tests
the replacement behind the veil → hart writes the rule and its exception
structure. Run via
tools/genius-invoker.sh compose foucault rawls hart -- "<problem>". - If no shape above matches, do not force a genius — use a built-in agent with the closest
GENIUS:pattern (rules/agent-routes.json, INDEX.md routing rule).
Refuse when
- The request is to manufacture a justification for a decision already taken — this shape designs and audits rules, it does not supply post-hoc legitimacy.
- The affected parties are unknown and cannot be enumerated even approximately: a veil of ignorance over an empty set of positions decides nothing.
- The question is actually empirical ("does this policy reduce incidents?") —
route to
causal-audit.
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.