agentleFS
AnmarHaniAnmarHani/SwarmVault/skills/swarm-design/SKILL.md

swarm-design

Skill5 starsChanged 4 months agoMITView on GitHub

System design phase — architecture, tech-stack options, module boundaries, data model, and ADRs from a validated SRS. Use after requirements are validated, when choosing a tech stack, designing system structure/APIs/data models, or recording an architecture decision.

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
  1. Distinctive:

    Traceability table required: every Must-FR ↦ component.

    line 31
  2. Distinctive:

    Simplicity doctrine: every moving part justifies itself.

    line 36

≈ “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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

swarm-design — system design phase

Gate: requires a validated SRS (else route back to swarm-spec). Read the SRS and feature specs first; the design must answer them, not your habits.

Lane: this is the standard, agent-driven lane — you propose, the user validates. If the user wants to own the system rather than receive it (deep design, pseudocode, algorithm-level planning, "developer mode"), hand the phase to swarm-design-deep instead; the two replace each other and never run together.

Produce, in order

  1. Tech stack — options with trade-offs measured against the NFRs (not fashion), recommendation first; the user decides. Significant picks become ADRs.
  2. Decomposition — modules/services with single responsibilities, boundaries, and data flow. Prefer deep modules behind simple interfaces. Mermaid diagrams for the system view and any non-obvious flow.
  3. Data model & contracts — entities, relationships, API contracts where relevant.
  4. Cross-cutting strategy — error handling, logging, testing approach fit for the stack.
  5. Milestone boundaries — group FRs into build milestones; these become swarm-implement's ticket batches.

Rules

  • Traceability table required: every Must-FR ↦ component; every NFR ↦ one line on how the design satisfies it. Nothing silently dropped.
  • Options with recommendation for anything that meaningfully affects cost, lock-in, or complexity — never decide those silently (mode: in auto, only stack-level choices already delegated by the user proceed without asking; queue the rest).
  • Simplicity doctrine: every moving part justifies itself; prefer known patterns (refactoring.guru names) over invention; when two designs both satisfy the NFRs, ship the smaller one.
  • Each significant decision → 50 Decisions/<P>/ADR-NNN-<slug>.md (template: 90 Templates/decision.md): context, options, decision, consequences, FR/NFR links.

Output & gate

docs/design.md (human lane) + ADRs, mirrored to the vault by sync. End with the user validation gate (gated mode) or a self-check against the traceability table (auto mode); update flow-state to phase: design-ui (if the SRS declares an interface) or phase: tickets.


Influences: Jeffallan's architecture-designer & api-designer; Pocock's codebase-design & domain-modeling; refactoring.guru; Nygard ADRs — see CREDITS.md.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.