Files
pbakaus_impeccable/scripts/lib/transformers/providers.js
T
Paul BakausandClaude Fable 5 fa1177ed9c Ship native subagent definitions for GitHub Copilot and Cursor
The github and cursor providers previously received only the generated
degraded/ inline fallbacks. Both harnesses support real custom subagents,
so the build now emits them from the same skill/agents/ source:

- GitHub Copilot: .github/agents/impeccable-<role>.agent.md with portable
  frontmatter only (name + description; omitting tools grants all tools,
  and Copilot has no documented model/effort/max-turns equivalents).
- Cursor: .cursor/agents/impeccable-<role>.md with name, description,
  model: inherit, is_background: false, and readonly derived from the
  agent's tool list (true only for the finish reviewer, which declares
  neither Write nor Edit). effort/max-turns are skipped because Cursor's
  effort option requires an explicit model id.

Agent bodies now also resolve {{scripts_path}} and strip rule markers in
the shared agentFormat pipeline, which fixes the previously unresolved
placeholder in the emitted Claude asset-producer agent.

The CLI installer places agents per scope: project installs write
<repo>/.github/agents/ and <repo>/.cursor/agents/; user-level installs
write ~/.copilot/agents/ (Copilot's user dir, not ~/.github/) and
~/.cursor/agents/, overwriting stale impeccable-* copies. Because
Copilot lets user-level agents shadow same-named project ones, a project
install warns when shadowing copies exist; Cursor gives project agents
precedence, so no warning there.

new-work.md and visualize.md extend their harness-naming clauses with
the Cursor and Copilot invocations. The degraded/ fallbacks keep
shipping for surfaces where the model still fails to delegate.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-28 16:16:57 -07:00

159 lines
5.8 KiB
JavaScript

/**
* Provider configurations for the transformer factory.
*
* Each config specifies:
* - provider: key into PROVIDER_PLACEHOLDERS (e.g. 'claude-code')
* - configDir: dot-directory name (e.g. '.claude')
* - displayName: human-readable name for log output (e.g. 'Claude Code')
* - providerTags: markdown block tags kept for this target (e.g. <codex>...</codex>)
* - frontmatterFields: which optional fields to emit beyond name + description
* - bodyTransform: optional function (body, skill) => transformed body
*/
export const PROVIDERS = {
cursor: {
provider: 'cursor',
providerTags: ['cursor'],
configDir: '.cursor',
displayName: 'Cursor',
frontmatterFields: ['license', 'compatibility', 'metadata'],
// Cursor subagents: `.cursor/agents/<name>.md` at repo level,
// `~/.cursor/agents/` at user level. Project agents take precedence over
// user ones, so installs simply overwrite on update.
agentFormat: 'cursor-md',
emitHooks: 'cursor',
// Cursor reads `.cursor/hooks.json`, not `.cursor/hooks/hooks.json`.
hooksManifestRel: 'hooks.json',
},
'claude-code': {
provider: 'claude-code',
providerTags: ['claude-code', 'claude'],
configDir: '.claude',
displayName: 'Claude Code',
frontmatterFields: ['user-invocable', 'argument-hint', 'license', 'compatibility', 'metadata', 'allowed-tools'],
agentFormat: 'claude-md',
emitHooks: 'claude',
// Project-local Claude Code hooks live in `.claude/settings.json`.
hooksManifestRel: 'settings.json',
},
gemini: {
provider: 'gemini',
providerTags: ['gemini'],
configDir: '.gemini',
displayName: 'Gemini',
frontmatterFields: [],
},
codex: {
provider: 'codex',
providerTags: ['codex'],
configDir: '.codex',
displayName: 'Codex',
frontmatterFields: [],
writeOpenAIMetadata: true,
// No agentFormat: the Codex subagent ships nested inside the skill's own
// agents/ folder (see CODEX_SKILL_PROVIDERS in factory.js), which Codex
// auto-discovers on install. No top-level .codex/agents/ sidecar is emitted.
emitHooks: 'codex',
// Codex discovers project-local hooks at `.codex/hooks.json`.
hooksManifestRel: 'hooks.json',
},
agents: {
provider: 'agents',
providerTags: ['agents', 'codex'],
configDir: '.agents',
displayName: 'Codex Repo Skills',
placeholderProvider: 'codex',
frontmatterFields: [],
writeOpenAIMetadata: true,
},
github: {
provider: 'github',
providerTags: ['github'],
configDir: '.github',
displayName: 'GitHub Copilot',
placeholderProvider: 'agents',
frontmatterFields: ['user-invocable', 'argument-hint', 'license', 'compatibility', 'metadata'],
// Copilot custom agents: `.github/agents/<name>.agent.md` at repo level,
// `~/.copilot/agents/` at user level (the CLI installer handles placement).
// The degraded/ fallbacks still ship for Copilot surfaces where the model
// fails to delegate; the .agent.md files are the real subagent path.
agentFormat: 'copilot-agent-md',
emitHooks: 'github',
// GitHub Copilot discovers repo-level hooks under `.github/hooks/*.json`.
hooksManifestRel: 'hooks/impeccable.json',
},
kiro: {
provider: 'kiro',
providerTags: ['kiro'],
configDir: '.kiro',
displayName: 'Kiro',
frontmatterFields: ['license', 'compatibility', 'metadata'],
},
opencode: {
provider: 'opencode',
providerTags: ['opencode'],
configDir: '.opencode',
displayName: 'OpenCode',
frontmatterFields: ['user-invocable', 'argument-hint', 'license', 'compatibility', 'metadata', 'allowed-tools'],
},
pi: {
provider: 'pi',
providerTags: ['pi'],
configDir: '.pi',
displayName: 'Pi',
frontmatterFields: ['license', 'compatibility', 'metadata', 'allowed-tools'],
},
qoder: {
provider: 'qoder',
providerTags: ['qoder'],
configDir: '.qoder',
displayName: 'Qoder',
frontmatterFields: ['user-invocable', 'argument-hint', 'license', 'compatibility', 'metadata', 'allowed-tools'],
},
'trae-cn': {
provider: 'trae-cn',
providerTags: ['trae-cn', 'trae'],
configDir: '.trae-cn',
displayName: 'Trae China',
placeholderProvider: 'trae',
frontmatterFields: ['user-invocable', 'argument-hint', 'license', 'compatibility', 'metadata'],
},
trae: {
provider: 'trae',
providerTags: ['trae'],
configDir: '.trae',
displayName: 'Trae',
frontmatterFields: ['user-invocable', 'argument-hint', 'license', 'compatibility', 'metadata'],
},
'rovo-dev': {
provider: 'rovo-dev',
providerTags: ['rovo-dev'],
configDir: '.rovodev',
displayName: 'Rovo Dev',
frontmatterFields: ['user-invocable', 'argument-hint', 'license', 'compatibility', 'metadata', 'allowed-tools'],
},
vibe: {
provider: 'vibe',
providerTags: ['vibe'],
configDir: '.vibe',
displayName: 'Mistral Vibe',
frontmatterFields: ['user-invocable', 'license', 'compatibility', 'metadata', 'allowed-tools'],
},
grok: {
provider: 'grok',
providerTags: ['grok'],
configDir: '.grok',
displayName: 'Grok Build',
// Grok's skill frontmatter matches the Agent Skills spec plus Claude-style
// extensions (user-invocable, argument-hint, allowed-tools, model, effort).
// See https://docs.x.ai/build/features/skills-plugins-marketplaces and
// ~/.grok/docs/user-guide/08-skills.md.
frontmatterFields: ['user-invocable', 'argument-hint', 'license', 'compatibility', 'metadata', 'allowed-tools'],
// Project/user agents are markdown with YAML frontmatter (Claude-compatible).
agentFormat: 'claude-md',
emitHooks: 'grok',
// Grok discovers project hooks from `.grok/hooks/*.json` (not a single
// settings.json). Claude tool-name matchers alias to Grok tools.
hooksManifestRel: 'hooks/impeccable.json',
},
};