Files
feedBack/.claude
Miguel_LZPFandBret Mogilefsky e4187d0054 feat: add cross-tool orientation, CI schema validation, and Claude Code surfaces
Adds the contributor- and AI-tool-facing infrastructure on top of the
modular docs from the previous commit. Lands AGENTS.md as the canonical
cross-tool orientation (read natively by Cursor, Copilot, Codex, Aider,
Cline, Continue, Cody, Devin, Replit Agent, and Claude Code), flips
CLAUDE.md to a 22-line pointer that uses Claude Code's @-import to
inline AGENTS.md, wires up plugin.json validation in CI, and adds the
Claude-specific automation surfaces under .claude/.

Cross-tool orientation:
  AGENTS.md (178 lines) — single source of truth: architecture, running
    the app, testing, git workflow, versioning, song formats, frontend
    and backend conventions, plugin authoring index, first-hour
    pitfalls, verification, house rules.
  CLAUDE.md (22 lines) — Claude Code memory file. Uses @AGENTS.md
    import (recursion depth 5) so the canonical content is inlined
    without duplication. Lists .claude/ surfaces.
  .github/copilot-instructions.md — Copilot custom instructions
    format; points at AGENTS.md and docs/PLUGIN_AUTHORING.md.
  .cursorrules — not added. Cursor reads AGENTS.md natively in 2026
    and .cursorrules is legacy.

Contribution hygiene (.github/):
  PULL_REQUEST_TEMPLATE.md — summary, linked issue, test plan, DCO
    and conventional-commit reminders. No AI-disclosure section.
  ISSUE_TEMPLATE/bug.yml — version, deployment, OS, plugins enabled,
    repro, logs (linked to docs/diagnostics-bundle-spec.md for
    redaction guidance).
  ISSUE_TEMPLATE/feature.yml — problem, proposed, alternatives,
    surface, plugin-author impact, license check.
  ISSUE_TEMPLATE/config.yml — disables blank issues; redirects
    plugin issues to plugin repos and security to the private
    advisory flow.

CI:
  .github/workflows/validate-plugins.yml — runs on changes to
    plugins/*, schema/, CONTRIBUTING.md, the test file, or the
    workflow itself. Installs jsonschema and pytest, validates every
    plugins/*/plugin.json against schema/plugin.schema.json, and runs
    the license-allowlist subset check.
  tests/test_plugin_schema.py — 8 parametrized tests: schema is
    well-formed, the 3 in-tree manifests validate, manifest id
    matches its parent directory name, schema license enum is a
    subset of CONTRIBUTING's curated allowlist.
  requirements-test.txt — append jsonschema>=4.0.
  .github/workflows/sync-version.yml — comment retargeted to
    AGENTS.md "Versioning" section.

Claude Code surfaces (.claude/):
  README.md — layout explanation. Spec-kit owns skills/speckit-*;
    repo-specific skills sit alongside. Hooks off by default;
    settings.json carries a commented opt-in example.
  skills/plugin-scaffold/SKILL.md — generates a plugin skeleton for
    type in {visualization, overlay, settings-only, routes-only}.
  skills/plugin-validate/SKILL.md — local pre-push check: validates
    plugin.json against schema, asserts declared files exist,
    enforces license allowlist.
  rules/plugin-author.md — globs scoped to plugins/**. Encodes the
    contracts from docs/PLUGIN_AUTHORING.md so AI suggestions don't
    drift from them (manifest required, context[\"log\"] over print,
    load_sibling over bare imports, playSong await discipline,
    settings.server_files conventions).
  agents/slopsmith-reviewer.md — plugin-aware reviewer subagent;
    invoke with @slopsmith-reviewer. 12-item checklist mirrors the
    rule and the schema.
  settings.json — empty hooks block plus a commented PostToolUse
    example for opt-in plugin.json validation on save.

Inbound-ref updates (files we own):
  README.md — \"AI Agent Guide\" points at AGENTS.md and notes
    .claude/ and copilot-instructions are tool-specific.
  CONTRIBUTING.md — \"Plugin System in CLAUDE.md\" -> docs/ and
    schema/; \"Git Workflow\" -> AGENTS.md#git-workflow.
  docs/sloppak-spec.md — plugin-system table cell -> docs/.

Out of scope (intentionally untouched):
  plugins/highway_3d/README.md (gitlink — plugin owns its docs).
  .specify/memory/constitution.md and other spec-kit artefacts
    (spec-kit owns that surface; CLAUDE.md still resolves
    transitively via the @-import).

Verification:
  pytest -q                                     # backend + schema tests pass
  python -c \"import json,glob,jsonschema; s=json.load(open('schema/plugin.schema.json')); [jsonschema.validate(json.load(open(p)), s) for p in sorted(glob.glob('plugins/*/plugin.json'))]\"
Signed-off-by: Miguel_LZPF <mgcdreamer@gmail.com>
2026-06-18 00:38:51 -07:00
..

.claude/ — Claude Code surfaces

This directory holds Claude Code artifacts: skills, rules, and subagents. Other AI tools (Cursor, Copilot, Codex, Aider) should read AGENTS.md instead — .claude/ is Claude-Code-specific.

Layout

.claude/
├── agents/                            Subagents (invoked via @<name>)
│   └── slopsmith-reviewer.md          Plugin-aware code review
├── rules/                             Globs-scoped behaviour rules
│   └── plugin-author.md               Rules that fire when editing plugins/**
├── skills/                            Skills (description triggers auto-invocation)
│   ├── plugin-scaffold/SKILL.md       Generate a new plugin skeleton
│   ├── plugin-validate/SKILL.md       Validate a plugin.json against the schema
│   └── speckit-*/                     Spec-kit skills (auto-generated; don't edit)
├── settings.json                      Repo defaults (no hooks enabled by default)
└── README.md                          You are here

Conventions

  • Spec-kit owns skills/speckit-*. Don't edit those manually — they're regenerated by the spec-kit installer (.specify/).
  • Repo skills live alongside spec-kit skills in skills/. New skills should have a kebab-case directory name and a SKILL.md with YAML frontmatter.
  • Hooks are off by default. Hooks in .claude/settings.json run in every contributor's Claude Code session — enabling one imposes work on everyone. The settings file ships with hooks: {} and a commented example so anyone who wants to opt in can copy it locally.
  • Subagent invocation. Use @slopsmith-reviewer to invoke a subagent explicitly. Plain auto-routing isn't currently configured.

Adding a new skill

  1. Create .claude/skills/<kebab-name>/SKILL.md.
  2. YAML frontmatter must include name (must match directory) and description (must read like a sentence — Claude Code matches the description against user intent for auto-invocation).
  3. Body: instructions Claude follows when the skill triggers. Keep it concrete and short — long skills bit-rot.

Adding a new rule

  1. Create .claude/rules/<topic>.md.
  2. Frontmatter may include globs: (array of glob patterns scoping when the rule applies). Without globs, the rule loads for every session.
  3. Body: terse, imperative guidance — "always do X", "never do Y", "if you see Z, …".

Why so little here?

The maintainer's personal Claude Code rules (context-mode routing, RTK, @git-worker, etc.) deliberately do not live in this repo. Those are user-global concerns in ~/.claude/ — keeping them out of the repo keeps the shared surface lean and makes the repo equally useful to contributors using Cursor, Copilot, plain editors, etc.