TrainDocs

You are running /evyasys:TrainDocs $ARGUMENTS.

Step 0 — Pre-flight

0a. Check .evyasys/docs/ for existing files: - No docs → proceed with full generation (all 37 documents). - Docs exist and $ARGUMENTS is empty → ask: "Docs already exist. Regenerate all, update specific, retrain, or abort?" - $ARGUMENTS is --update → confirm: "Regenerate all 37 documents?" - $ARGUMENTS is --update <filename> → confirm: "Regenerate .evyasys/docs/<filename> only?" - $ARGUMENTS is --retrain → switch to retrain mode (see PROMPT.md Step 0-R): read last generation date from INDEX.md, detect changed areas via git log, regenerate only affected docs. Note: 37 documents total — 35 core + fe/ACCESSIBILITY.md + fe/VISUAL_QUALITY.md (skip if no frontend layer). 0b. Read CLAUDE.md from project root if present. Carry its content into ARCHITECTURE.md and RULES.md. 0c. Find the plugin's installed workflow directory, then load AGENT.md, PROMPT.md, and DOC_MANIFEST.md from it:

macOS / Linux (Bash): `EVYA_AI=$(find "$HOME/.claude/plugins" -maxdepth 6 -type d -name ".ai" 2>/dev/null | grep -i "EvyaGovernance" | head -1); [ -z "$EVYA_AI" ] && EVYA_AI=".ai"; echo "$EVYA_AI"`

Windows (PowerShell): `$EVYA_AI = (Get-ChildItem "$env:USERPROFILE\.claude\plugins" -Recurse -Directory -Filter ".ai" -ErrorAction SilentlyContinue | Where-Object { $_.FullName -like '*EvyaGovernance*' } | Select-Object -First 1 -ExpandProperty FullName); if (-not $EVYA_AI) { $EVYA_AI = ".ai" }; Write-Output $EVYA_AI`

Load `<plugin-ai>/workflows/create-docs/AGENT.md`, `<plugin-ai>/workflows/create-docs/PROMPT.md`, and `<plugin-ai>/workflows/create-docs/DOC_MANIFEST.md` (replace `<plugin-ai>` with the printed path). If any file is not found at the plugin path, fall back to the same filename under `.ai/workflows/create-docs/`.

Phase 1 — Project scan (complete before writing any document)

1. Tech stack: read package.json, requirements.txt, pyproject.toml, Cargo.toml, go.mod, pom.xml, tsconfig.json, docker-compose.yml, Dockerfile*.
2. Source structure: Glob src/**, app/**, lib/**, components/**, pages/**, api/**, services/**, models/**, repositories/** — map folder layout and naming.
3. Tooling: read .eslintrc*, .prettierrc*, .editorconfig, jest.config.*, vitest.config.*, playwright.config.*, tailwind.config.*, prisma/schema.prisma.
4. CI/CD: read .github/workflows/**, azure-pipelines.yml, Jenkinsfile.
5. Code sampling: read 3–5 representative source files per layer to infer patterns, naming, and conventions.
6. Existing docs: read README.md, CONTRIBUTING.md, any files under docs/.