agentleFS
Sign inStart free
ShiningPlagueShiningPlague/shiningplague-studio/skills/architecture-decision/SKILL.md

architecture-decision

Skill0 starsChanged 4 months agoMITView on GitHub

Use when locking a technical or design decision into an ADR. Fires after /brainstorming or /design-review when decisions need to be formally recorded. Batch 2-3 decisions per ADR. ShiningPlague-adopted: full implementation with studio ADR template (Live Sources table), technical-director dispatch with TD-ADR gate, registry + control-manifest + tr-registry update pipeline.

32 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use component-release, soak-test and 1 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • Microsoft AzureSkill

    component-release

    Prepares a verifiable Agent Landing Zone UI release. Use for release branches, VERSION, changelog entries, tags, release notes, and compatibility evidence; never publish without explicit human approval.

    Also in 4 projects that use this

  • lilangMaxSkill

    soak-test

    Generate a soak test protocol for extended play sessions. Defines what to observe, measure, and log during long play sessions to surface slow leaks, fatigue effects, and edge cases that only appear after sustained play. Primarily used in Polish and Release phases.

    Also in 7 projects that use this

  • lilangMaxSkill

    team-live-ops

    Orchestrate the live-ops team for post-launch content planning: coordinates live-ops-designer, economy-designer, analytics-engineer, community-manager, writer, and narrative-director to design and plan a season, event, or live content update.

    Also in 7 projects that use this

The file

Show its rules and text

We split this skill into the rules it gives an agent

13 rules. 0 are common, 3 are distinctive, and 10 are very specific to this project.

  • 0

    Common

    Most of the 30 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 10

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    After a brainstorm spec reaches FINAL with decisions that constrain implementation

    line 29
  2. Very specific:

    After /design-review APPROVE when the design locks choices

    line 30
  3. Distinctive:

    When the team says "we need to decide X before building"

    line 31
  4. Very specific:

    Read docs/adr/TEMPLATE.md for the format

    line 42
  5. Very specific:

    Read existing ADRs at docs/adr/

    line 43
  6. Distinctive:

    Read the source spec for decisions being locked

    line 44
  7. Very specific:

    Read docs/architecture/control-manifest.md to check for conflicts

    line 45
  8. Very specific:

    Update docs/architecture/control-manifest.md with new programmer rules

    line 66
  9. Very specific:

    Update data/_schemas/system_registry.json — add ADR to relevant system entries' decisions[] and to documentation_stack.adr_index.adrs[]

    line 67
  10. Very specific:

    Update docs/architecture/tr-registry.yaml — mark covered requirements

    line 68
  11. Very specific:

    After Accept → propose /writing-plans to plan implementation.

    line 74
  12. Very specific:

    After Accept → propose /create-control-manifest if many rules changed.

    line 75
  13. Very specific:

    After Accept → propose /verification-before-completion before any "ADR shipped" claim.

    line 76

“Common” is measured: kept by at least 6 in 10 of the 30 edited versions of this file we could read. “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

Architecture Decision

🌱 ShiningPlague-adopted. Originally Donchitos (barebones stub); battle-tested on a shipped Godot project. Upstream was a pointer-only file; the studio version is the full implementation — ADR authoring with Live Sources, batch rule, director gate, downstream updates.

Guided ADR authoring using the studio + Donchitos merged template at docs/adr/TEMPLATE.md.

If an artifact named here is absent: say so plainly in one line, skip that step, and continue. Never invent the file to satisfy a checklist, and never fail a close because an optional artifact was never created.

When to Fire

  • After a brainstorm spec reaches FINAL with decisions that constrain implementation
  • After /design-review APPROVE when the design locks choices
  • When the team says "we need to decide X before building"

Inputs

  • Decisions to lock — 2-3 related decisions per ADR (per Step sizing retrospective)
  • Source spec — the brainstorm spec that generated these decisions
  • Existing ADRs — docs/adr/ for cross-reference

Procedure

1. Read Context

  • Read docs/adr/TEMPLATE.md for the format
  • Read existing ADRs at docs/adr/
  • Read the source spec for decisions being locked
  • Read docs/architecture/control-manifest.md to check for conflicts

2. Assign ADR Number

Next number = highest existing + 1. Naming: docs/adr/NNN-<slug>.md.

3. Draft ADR Sections

Using the template (Donchitos + studio merged format with Live Sources):

  • Title + Status (Proposed)
  • Context — what problem, what constraints
  • Decision — what we chose and why
  • Consequences — positive + negative + trade-offs
  • Alternatives Considered — what was rejected and why
  • Live Sources — which JSON/data files are authoritative for this decision's values

4. Dispatch Technical Director

Spawn technical-director with gate TD-ADR from .claude/docs/director-gates.md. Pass: draft ADR, related existing ADRs, engine constraints. Await verdict.

5. If APPROVE → Accept

  • Set status to Accepted
  • Update docs/architecture/control-manifest.md with new programmer rules
  • Update data/_schemas/system_registry.json — add ADR to relevant system entries' decisions[] and to documentation_stack.adr_index.adrs[]
  • Update docs/architecture/tr-registry.yaml — mark covered requirements

6. Present to Designer

Show the ADR for final approval before committing.

Chain-Propose

After Accept → propose /writing-plans to plan implementation. After Accept → propose /create-control-manifest if many rules changed. After Accept → propose /verification-before-completion before any "ADR shipped" claim.

Batch rule

One ADR covers 2–3 related decisions from one design phase. Don't write one ADR per decision. Illustrative example — these filenames are not shipped, they show the shape:

  • 001-your-system-schema-and-slots.md would cover the data schema + the capacity model + the stacking policy for one system.
  • 002-your-system-pipeline-and-effects.md (later) would cover table resolution + effect architecture + the autoload extraction.

Project Paths

  • ADRs: docs/adr/NNN-<slug>.md (docs/architecture/ holds the blueprint + manifest, never ADRs)
  • Template: docs/adr/TEMPLATE.md
  • Control manifest: docs/architecture/control-manifest.md
  • TR registry: docs/architecture/tr-registry.yaml
  • Director gates: .claude/docs/director-gates.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.

Check your setup, free