boundary-design
Skill8 starsChanged 58 days agoMITView on GitHub
Draw the line where it costs least. Use for build-vs-buy calls, "where does this module/team/service boundary belong?", APIs and abstractions forcing users into implementation vocabulary, tools that should augment rather than automate, hardcoded decisions that belong at runtime, or information nobody can find.
In this repository16 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- causal-auditSkill
- decision-bias-checkSkill
- estimationSkill
- evidence-synthesisSkill
- experience-and-transmissionSkill
- failure-forensicsSkill
- formal-correctnessSkill
- measurement-disciplineSkill
- narrative-sensemakingSkill
- normative-designSkill
- 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 43
If no shape above matches, use a built-in agent with the closest GENIUS: pattern from rules/agent-routes.json instead.
≈ “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
Boundary Design
Problem shape: something must be split, wrapped, or interfaced — a system into modules, a workflow between human and tool, a product between build and buy — and the current boundary leaks implementation detail, transaction cost, or cognitive load across it.
Relevant geniuses
| Agent | Use when |
| coase | build vs buy; team/service boundary placement — compare transaction costs across the boundary, not ideology |
| hopper | users forced to think in implementation vocabulary — build the translator; debugging under-invested; a tool defended out of familiarity |
| engelbart | "automate this" proposed where augmenting the person is the real win; the team doesn't use its own tool; design for expert ceiling, not just novice floor |
| kay | decisions hardcoded that could bind late; tight coupling via direct calls; users need to modify the system at runtime |
| alexander | decompose by misfit; the design needs a generative sequence of patterns rather than a top-down blueprint |
| simon | the system should be nearly decomposable — strong interactions inside modules, weak between; hierarchy as the default shape |
| liskov | the boundary is an interface — write the behavioral contract so implementations are substitutable |
| ranganathan | information exists but nobody can find it — faceted classification and findability laws |
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 and answer in its<output-format>. Every proposed boundary states what crosses it, what it hides, and what it costs (Clean Architecture layer rules apply, §2.2). - Typical chain: coase places the boundary → simon shapes the decomposition →
liskov contracts the interface. Run via
tools/genius-invoker.sh compose coase liskov -- "<problem>". - If no shape above matches, use a built-in agent with the closest
GENIUS:pattern fromrules/agent-routes.jsoninstead.
Refuse when
- The boundary is being drawn to match an org chart against the dependency evidence — surface the conflict instead of ratifying it.
- Fewer than three concrete uses exist for a proposed abstraction (coding-standards §3.3: premature abstraction is worse than duplication).
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.