* feat: generate commented wmill.yaml template and add config reference command Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: add missing options to config reference (promotion, skipBranchValidation, commonSpecificItems) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * refactor: generate YAML template from CONFIG_REFERENCE instead of handwritten string Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: preserve YAML comments when binding workspace profile during init Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * refactor: simplify to `wmill config` and reorder table columns Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * feat: generate JSON Schema for wmill.yaml editor autocomplete and validation Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * refactor: remove redundant templateValue fields and make specificItemsSchema data-driven Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * refactor: use native JSON Schema types in CONFIG_REFERENCE, strip non-schema keys for generation Eliminates typeToJsonSchema, specificItemsSchema, codebaseItemSchema, branchConfigSchema, and the complex generateJsonSchema body. Each CONFIG_REFERENCE entry is now a JSON Schema property with extra metadata. Schema generation just iterates and strips non-schema keys. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * refactor: remove typeLabel and displayType — use schema types directly Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * refactor: remove hidden entries, auto-expand nested schemas in reference table Sub-fields (codebases[], gitBranches.<branch>.*) are now derived from the parent's inline schema instead of being maintained as duplicate hidden entries. Removes 29 entries and the hidden field entirely. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: use console.log for JSON output and quote YAML-special branch names Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * chore: regenerate system prompts to include new config command Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: review feedback + add tests for template, schema, and config reference - Use console.log for --json output (no ANSI escape codes) - Quote branch names with YAML-special characters - Add 28 tests covering template generation, JSON Schema validation, config reference formatting, and CONFIG_REFERENCE integrity Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * feat: add generate-schema script and commit wmill.schema.json to repo Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * refactor: remove schema.json generation from wmill init Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: eliminate read-back cycle, harden yamlKey, fix triple negation Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
System Prompts
This directory contains the single source of truth for AI system prompts used by both the frontend copilot and CLI guidance.
Structure
system_prompts/
├── base/ # Core instruction templates (manually written)
├── languages/ # Language-specific instructions (manually written)
└── auto-generated/ # Auto-generated files (DO NOT EDIT)
├── sdks/ # SDK documentation
├── cli/ # CLI command documentation
├── prompts.ts # TypeScript exports
└── index.ts # Helper functions
Usage
Regenerating Prompts
When SDK methods or the OpenFlow schema change, run:
python system_prompts/generate.py
This will:
- Parse TypeScript and Python SDK files to extract function signatures
- Parse the OpenFlow YAML schema
- Parse the CLI commands
- Assemble complete prompts from markdown files
- Generate TypeScript exports in
auto-generated/
Scope
These system prompts contain ONLY:
- How to write Windmill scripts (language syntax, conventions, SDK usage)
- How to structure Windmill flows (OpenFlow schema, module types, data flow)
- Resource type handling, S3 operations
They DO NOT contain:
- Tool usage instructions (edit_code, set_flow_json, etc.)
- IDE/editor specific commands
- Testing tool invocations
Tool instructions are added separately by the frontend and CLI.
Integration
Frontend
Uses Vite path alias $system_prompts pointing to auto-generated/:
import { FLOW_GUIDANCE } from "$system_prompts/flow";
import { getLangContext } from "$system_prompts/languages";
CLI
Generates /cli/src/guidance/skills.ts with embedded skill content for wmill init.
Editing Guidelines
- Edit markdown files in
base/,languages/ - Never edit files in
auto-generated/directly - After editing, run
generate.pyto update exports