mirror of
https://github.com/pbakaus/impeccable.git
synced 2026-09-14 23:26:39 +03:00
Migrate to unified skills architecture, add Copilot & Antigravity (v2.0.0)
All commands are now skills with user-invokable: true. Source lives in
source/skills/{name}/SKILL.md. Added VS Code Copilot (.agents/skills/)
and Google Antigravity (.agent/skills/) transformers. All 6 providers
output to skills directories only — no more commands/prompts dirs.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
Paul Bakaus
co-authored by
Claude Opus 4.6
Claude Opus 4.6
parent
a94e237359
commit
87106b5271
@@ -2,110 +2,56 @@ import path from 'path';
|
||||
import { cleanDir, ensureDir, writeFile, generateYamlFrontmatter, replacePlaceholders } from '../utils.js';
|
||||
|
||||
/**
|
||||
* Generate markdown from structured patterns/antipatterns data
|
||||
*/
|
||||
function generatePatternsMarkdown(patterns) {
|
||||
if (!patterns || (!patterns.patterns?.length && !patterns.antipatterns?.length)) {
|
||||
return '';
|
||||
}
|
||||
|
||||
let md = `## Design Patterns Reference
|
||||
|
||||
This reference defines what TO do and what NOT to do when creating frontend interfaces. These patterns fight against model bias—the tendency of LLMs to converge on the same predictable choices.
|
||||
|
||||
### What TO Do (Patterns)
|
||||
|
||||
Focus on intentional, distinctive design choices:
|
||||
`;
|
||||
|
||||
for (const category of patterns.patterns || []) {
|
||||
md += `\n**${category.name}**:\n`;
|
||||
for (const item of category.items || []) {
|
||||
md += `- ${item}\n`;
|
||||
}
|
||||
}
|
||||
|
||||
md += `
|
||||
### What NOT to Do (Anti-Patterns)
|
||||
|
||||
These patterns create generic "AI slop" aesthetics:
|
||||
`;
|
||||
|
||||
for (const category of patterns.antipatterns || []) {
|
||||
md += `\n**${category.name}**:\n`;
|
||||
for (const item of category.items || []) {
|
||||
md += `- ${item}\n`;
|
||||
}
|
||||
}
|
||||
|
||||
md += `
|
||||
These anti-patterns are baked into training data from countless generic templates. Without explicit guidance, AI reproduces them. This skill ensures your AI knows both what to do AND what to avoid.
|
||||
`;
|
||||
|
||||
return md;
|
||||
}
|
||||
|
||||
/**
|
||||
* Codex Transformer (Full Featured - Agent Skills Standard)
|
||||
* Codex Transformer (Skills Only)
|
||||
*
|
||||
* Commands: Uses argument-hint format with $VARIABLE placeholders in .codex/prompts/
|
||||
* Skills: Uses Agent Skills standard with SKILL.md in .codex/skills/{name}/
|
||||
* Reference files are copied to skill subdirectories
|
||||
* All skills output to .codex/skills/{name}/SKILL.md
|
||||
* Frontmatter: name, description, argument-hint (from args for user-invokable)
|
||||
* For user-invokable skills: {{argname}} becomes $ARGNAME in body
|
||||
*
|
||||
* @param {Array} skills - All skills (including user-invokable ones)
|
||||
* @param {string} distDir - Distribution output directory
|
||||
* @param {Object} patterns - Design patterns data (unused)
|
||||
* @param {Object} options - Optional settings
|
||||
* @param {string} options.prefix - Prefix to add to command names (e.g., 'i-')
|
||||
* @param {string} options.prefix - Prefix to add to user-invokable skill names (e.g., 'i-')
|
||||
* @param {string} options.outputSuffix - Suffix for output directory (e.g., '-prefixed')
|
||||
*/
|
||||
export function transformCodex(commands, skills, distDir, patterns = null, options = {}) {
|
||||
export function transformCodex(skills, distDir, patterns = null, options = {}) {
|
||||
const { prefix = '', outputSuffix = '' } = options;
|
||||
const codexDir = path.join(distDir, `codex${outputSuffix}`);
|
||||
const promptsDir = path.join(codexDir, '.codex/prompts');
|
||||
const skillsDir = path.join(codexDir, '.codex/skills');
|
||||
|
||||
cleanDir(codexDir);
|
||||
ensureDir(promptsDir);
|
||||
ensureDir(skillsDir);
|
||||
|
||||
// Commands: Transform to Codex prompt format
|
||||
for (const command of commands) {
|
||||
const commandName = `${prefix}${command.name}`;
|
||||
const yamlLines = ['---'];
|
||||
yamlLines.push(`description: ${command.description}`);
|
||||
|
||||
// Build argument-hint from args array
|
||||
if (command.args && command.args.length > 0) {
|
||||
const hints = command.args.map(arg => {
|
||||
const hint = arg.required ? `<${arg.name}>` : `[${arg.name.toUpperCase()}=<value>]`;
|
||||
return hint;
|
||||
});
|
||||
yamlLines.push(`argument-hint: ${hints.join(' ')}`);
|
||||
}
|
||||
|
||||
yamlLines.push('---');
|
||||
|
||||
// First replace our placeholders, then transform remaining {{argname}} to $ARGNAME
|
||||
let body = replacePlaceholders(command.body, 'codex');
|
||||
body = body.replace(/\{\{([^}]+)\}\}/g, (match, argName) => {
|
||||
return `$${argName.toUpperCase()}`;
|
||||
});
|
||||
|
||||
const content = `${yamlLines.join('\n')}\n\n${body}`;
|
||||
const outputPath = path.join(promptsDir, `${commandName}.md`);
|
||||
writeFile(outputPath, content);
|
||||
}
|
||||
|
||||
// Skills: Use Agent Skills standard with SKILL.md in subdirectories
|
||||
let refCount = 0;
|
||||
for (const skill of skills) {
|
||||
const skillDir = path.join(skillsDir, skill.name);
|
||||
const skillName = skill.userInvokable ? `${prefix}${skill.name}` : skill.name;
|
||||
const skillDir = path.join(skillsDir, skillName);
|
||||
|
||||
const frontmatter = generateYamlFrontmatter({
|
||||
name: skill.name,
|
||||
const frontmatterObj = {
|
||||
name: skillName,
|
||||
description: skill.description,
|
||||
...(skill.license && { license: skill.license })
|
||||
});
|
||||
};
|
||||
|
||||
// Build argument-hint from args array for user-invokable skills
|
||||
if (skill.userInvokable && skill.args && skill.args.length > 0) {
|
||||
const hints = skill.args.map(arg => {
|
||||
return arg.required ? `<${arg.name}>` : `[${arg.name.toUpperCase()}=<value>]`;
|
||||
});
|
||||
frontmatterObj['argument-hint'] = hints.join(' ');
|
||||
}
|
||||
if (skill.license) frontmatterObj.license = skill.license;
|
||||
|
||||
const frontmatter = generateYamlFrontmatter(frontmatterObj);
|
||||
|
||||
let skillBody = replacePlaceholders(skill.body, 'codex');
|
||||
// For user-invokable skills, transform remaining {{argname}} to $ARGNAME
|
||||
if (skill.userInvokable) {
|
||||
skillBody = skillBody.replace(/\{\{([^}]+)\}\}/g, (match, argName) => {
|
||||
return `$${argName.toUpperCase()}`;
|
||||
});
|
||||
}
|
||||
|
||||
const skillBody = replacePlaceholders(skill.body, 'codex');
|
||||
const content = `${frontmatter}\n\n${skillBody}`;
|
||||
const outputPath = path.join(skillDir, 'SKILL.md');
|
||||
writeFile(outputPath, content);
|
||||
@@ -123,7 +69,8 @@ export function transformCodex(commands, skills, distDir, patterns = null, optio
|
||||
}
|
||||
}
|
||||
|
||||
const userInvokableCount = skills.filter(s => s.userInvokable).length;
|
||||
const refInfo = refCount > 0 ? ` (${refCount} reference files)` : '';
|
||||
const prefixInfo = prefix ? ` [${prefix}prefixed]` : '';
|
||||
console.log(`✓ Codex${prefixInfo}: ${commands.length} prompts, ${skills.length} skills${refInfo}`);
|
||||
console.log(`✓ Codex${prefixInfo}: ${skills.length} skills (${userInvokableCount} user-invokable)${refInfo}`);
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user