@aragon-agent/cli 0.6.29 → 0.6.32

package/skills/project-indexer/templates/config-template.md

@@ -1,69 +1,69 @@
-# Index Configuration
-
-> Configuration for project-indexer. Edit this file to customize index generation.
-
-## Excluded Directories
-
-Directories to skip during indexing:
-
-- node_modules
-- .git
-- dist
-- build
-- __pycache__
-- .venv
-- venv
-- vendor
-- coverage
-- .next
-- .nuxt
-- .cache
-- target
-- bin
-- obj
-- .idea
-- .vscode
-
-## Priority Directories
-
-Directories to index in detail (leave empty for auto-detection):
-
-- src/
-- lib/
-- app/
-- core/
-
-## Tech Stack
-
-Detected or specified technology stack:
-
-- **Frontend:** {React | Vue | Angular | Svelte | Next.js | Nuxt | etc.}
-- **Backend:** {Express | FastAPI | Django | Rails | Spring | etc.}
-- **Language:** {TypeScript | JavaScript | Python | Go | etc.}
-
-## Index Settings
-
-- **Generated:** {YYYY-MM-DD HH:MM:SS}
-- **Project Root:** {/path/to/project}
-- **Index Version:** 1.0
-- **Total Files Indexed:** {N}
-- **Total Symbols Extracted:** {N}
-
-## Clean Code Settings
-
-- Enabled: {Yes | No}
-- Preset: {strict | relaxed | custom | disabled}
-- MAX_FILE_LINES: 1000
-- MAX_METHOD_LINES: 60
-- MAX_FUNCTION_PARAMS: 5
-- MAX_CYCLOMATIC: 10
-- MAX_LINE_LENGTH: 100
-- MAX_NESTING_DEPTH: 4
-
-## Regeneration Notes
-
-Notes for future regeneration:
-
-- {Any special considerations}
-- {Files that need manual review}
-- {Known limitations}
+# Index Configuration
+
+> Configuration for project-indexer. Edit this file to customize index generation.
+
+## Excluded Directories
+
+Directories to skip during indexing:
+
+- node_modules
+- .git
+- dist
+- build
+- __pycache__
+- .venv
+- venv
+- vendor
+- coverage
+- .next
+- .nuxt
+- .cache
+- target
+- bin
+- obj
+- .idea
+- .vscode
+
+## Priority Directories
+
+Directories to index in detail (leave empty for auto-detection):
+
+- src/
+- lib/
+- app/
+- core/
+
+## Tech Stack
+
+Detected or specified technology stack:
+
+- **Frontend:** {React | Vue | Angular | Svelte | Next.js | Nuxt | etc.}
+- **Backend:** {Express | FastAPI | Django | Rails | Spring | etc.}
+- **Language:** {TypeScript | JavaScript | Python | Go | etc.}
+
+## Index Settings
+
+- **Generated:** {YYYY-MM-DD HH:MM:SS}
+- **Project Root:** {/path/to/project}
+- **Index Version:** 1.0
+- **Total Files Indexed:** {N}
+- **Total Symbols Extracted:** {N}
+
+## Clean Code Settings
+
+- Enabled: {Yes | No}
+- Preset: {strict | relaxed | custom | disabled}
+- MAX_FILE_LINES: 1000
+- MAX_METHOD_LINES: 60
+- MAX_FUNCTION_PARAMS: 5
+- MAX_CYCLOMATIC: 10
+- MAX_LINE_LENGTH: 100
+- MAX_NESTING_DEPTH: 4
+
+## Regeneration Notes
+
+Notes for future regeneration:
+
+- {Any special considerations}
+- {Files that need manual review}
+- {Known limitations}

package/skills/project-indexer/templates/index-template.md

@@ -1,129 +1,129 @@
-# Project Index: {project-name}
-
-> Auto-generated by project-indexer | Last updated: {YYYY-MM-DD}
-
-## Project Overview
-
-- **Type:** {Frontend | Backend | Fullstack | Library | CLI | Monorepo}
-- **Languages:** {TypeScript (70%), Python (30%)}
-- **Frameworks:** {Next.js, FastAPI, etc.}
-- **Entry Points:** `{src/index.ts}`, `{api/main.py}`
-
----
-
-## Feature Map
-
-### {Feature Name} (`{path}/`)
-
-Entry: `{entry-file}`
-
-{One-line description of this feature area}
-
-**Key exports:**
-- `{functionName}({params}): {returnType}` - {description}
-- `{ClassName}` - {description}
-- `{CONSTANT_NAME}` - {description}
-
----
-
-### Example: Authentication (`src/auth/`)
-
-Entry: `src/auth/index.ts`
-
-User authentication, session management, and authorization logic.
-
-**Key exports:**
-- `login(credentials: LoginInput): Promise<User>` - Authenticate user with email/password
-- `logout(): void` - Clear session and revoke tokens
-- `useAuth(): AuthContext` - React hook for accessing auth state
-- `withAuth(Component): Component` - HOC for protected routes
-- `AuthProvider` - Context provider for auth state
-
----
-
-### Example: API Layer (`src/api/`)
-
-Entry: `src/api/client.ts`
-
-REST API client with error handling, request interceptors, and response caching.
-
-**Key exports:**
-- `apiClient` - Pre-configured axios instance
-- `useQuery<T>(endpoint: string): SWRResponse<T>` - Data fetching hook with caching
-- `useMutation<T>(endpoint: string): MutationResult<T>` - Data mutation hook
-- `ApiError` - Custom error class for API errors
-
----
-
-## Module Dependencies
-
-```
-{module-a} → {module-b} ({relationship description})
-```
-
-Example:
-```
-auth → api (uses apiClient for authentication requests)
-pages → auth, components (imports hooks and UI components)
-api → (standalone, no internal dependencies)
-components → utils (uses helper functions)
-hooks → api, auth (combines data fetching with auth state)
-```
-
----
-
-## File Index
-
-### {directory}/
-
-| File | Description |
-|------|-------------|
-| {filename.ext} | {One-line description} |
-
----
-
-### Example: src/auth/
-
-| File | Description |
-|------|-------------|
-| index.ts | Module entry point, re-exports public API |
-| login.ts | Login logic, credential validation, token handling |
-| logout.ts | Logout logic, session cleanup, token revocation |
-| session.ts | Session storage, refresh token management |
-| hooks.ts | React hooks: useAuth, useUser, usePermissions |
-| context.tsx | AuthContext and AuthProvider implementation |
-| types.ts | TypeScript types for auth entities |
-| constants.ts | Auth-related constants and configuration |
-
-### Example: src/api/
-
-| File | Description |
-|------|-------------|
-| client.ts | Axios instance with interceptors and base config |
-| endpoints.ts | API endpoint path constants |
-| types.ts | Request/response type definitions |
-| errors.ts | Custom error classes and error handling utilities |
-| hooks.ts | Data fetching hooks: useQuery, useMutation |
-| cache.ts | Response caching configuration |
-
-### Example: src/components/
-
-| File | Description |
-|------|-------------|
-| Button.tsx | Reusable button with variants (primary, secondary, ghost) |
-| Modal.tsx | Dialog component with portal and focus trap |
-| Toast.tsx | Notification toast with auto-dismiss |
-| Spinner.tsx | Loading spinner component |
-| Form/index.tsx | Form wrapper with validation context |
-| Form/Input.tsx | Text input with label and error display |
-| Form/Select.tsx | Dropdown select component |
-| Form/Checkbox.tsx | Checkbox with label |
-| Layout/Header.tsx | App header with navigation |
-| Layout/Sidebar.tsx | Collapsible sidebar navigation |
-| Layout/Footer.tsx | App footer |
-
----
-
-## Notes
-
-{Any additional notes about the project structure, conventions, or important information}
+# Project Index: {project-name}
+
+> Auto-generated by project-indexer | Last updated: {YYYY-MM-DD}
+
+## Project Overview
+
+- **Type:** {Frontend | Backend | Fullstack | Library | CLI | Monorepo}
+- **Languages:** {TypeScript (70%), Python (30%)}
+- **Frameworks:** {Next.js, FastAPI, etc.}
+- **Entry Points:** `{src/index.ts}`, `{api/main.py}`
+
+---
+
+## Feature Map
+
+### {Feature Name} (`{path}/`)
+
+Entry: `{entry-file}`
+
+{One-line description of this feature area}
+
+**Key exports:**
+- `{functionName}({params}): {returnType}` - {description}
+- `{ClassName}` - {description}
+- `{CONSTANT_NAME}` - {description}
+
+---
+
+### Example: Authentication (`src/auth/`)
+
+Entry: `src/auth/index.ts`
+
+User authentication, session management, and authorization logic.
+
+**Key exports:**
+- `login(credentials: LoginInput): Promise<User>` - Authenticate user with email/password
+- `logout(): void` - Clear session and revoke tokens
+- `useAuth(): AuthContext` - React hook for accessing auth state
+- `withAuth(Component): Component` - HOC for protected routes
+- `AuthProvider` - Context provider for auth state
+
+---
+
+### Example: API Layer (`src/api/`)
+
+Entry: `src/api/client.ts`
+
+REST API client with error handling, request interceptors, and response caching.
+
+**Key exports:**
+- `apiClient` - Pre-configured axios instance
+- `useQuery<T>(endpoint: string): SWRResponse<T>` - Data fetching hook with caching
+- `useMutation<T>(endpoint: string): MutationResult<T>` - Data mutation hook
+- `ApiError` - Custom error class for API errors
+
+---
+
+## Module Dependencies
+
+```
+{module-a} → {module-b} ({relationship description})
+```
+
+Example:
+```
+auth → api (uses apiClient for authentication requests)
+pages → auth, components (imports hooks and UI components)
+api → (standalone, no internal dependencies)
+components → utils (uses helper functions)
+hooks → api, auth (combines data fetching with auth state)
+```
+
+---
+
+## File Index
+
+### {directory}/
+
+| File | Description |
+|------|-------------|
+| {filename.ext} | {One-line description} |
+
+---
+
+### Example: src/auth/
+
+| File | Description |
+|------|-------------|
+| index.ts | Module entry point, re-exports public API |
+| login.ts | Login logic, credential validation, token handling |
+| logout.ts | Logout logic, session cleanup, token revocation |
+| session.ts | Session storage, refresh token management |
+| hooks.ts | React hooks: useAuth, useUser, usePermissions |
+| context.tsx | AuthContext and AuthProvider implementation |
+| types.ts | TypeScript types for auth entities |
+| constants.ts | Auth-related constants and configuration |
+
+### Example: src/api/
+
+| File | Description |
+|------|-------------|
+| client.ts | Axios instance with interceptors and base config |
+| endpoints.ts | API endpoint path constants |
+| types.ts | Request/response type definitions |
+| errors.ts | Custom error classes and error handling utilities |
+| hooks.ts | Data fetching hooks: useQuery, useMutation |
+| cache.ts | Response caching configuration |
+
+### Example: src/components/
+
+| File | Description |
+|------|-------------|
+| Button.tsx | Reusable button with variants (primary, secondary, ghost) |
+| Modal.tsx | Dialog component with portal and focus trap |
+| Toast.tsx | Notification toast with auto-dismiss |
+| Spinner.tsx | Loading spinner component |
+| Form/index.tsx | Form wrapper with validation context |
+| Form/Input.tsx | Text input with label and error display |
+| Form/Select.tsx | Dropdown select component |
+| Form/Checkbox.tsx | Checkbox with label |
+| Layout/Header.tsx | App header with navigation |
+| Layout/Sidebar.tsx | Collapsible sidebar navigation |
+| Layout/Footer.tsx | App footer |
+
+---
+
+## Notes
+
+{Any additional notes about the project structure, conventions, or important information}

package/skills/skill-creator/SKILL.md

@@ -1,96 +1,96 @@
----
-name: skill-creator
-description: Author a new AragonAgent skill. Use when the user asks to save a procedure as a skill, or when you notice you have re-derived the same multi-step process twice in a session.
-version: 1.1.0
-license: MIT
-activation: auto
-keywords: [meta, authoring, skills]
-# Enforced as a per-turn tool ceiling. `edit_file` is here because revising an
-# existing skill is the single most common thing this skill is asked to do —
-# without it, "fix the description on my deploy skill" fails on the first edit.
-allowed-tools:
-  - read_file
-  - write_file
-  - edit_file
-  - list_dir
-  - skill_create
----
-
-# Skill Creator
-
-Write a skill that another model instance will actually reach for at the right
-moment. Most bad skills are not badly written; they are badly *described* and
-badly *scoped*.
-
-## When to use
-
-- The user says "remember how to do this", "save this as a skill", or similar.
-- You have worked out a non-obvious, repeatable procedure and expect to need it
-  again.
-- You notice you have rediscovered the same steps twice in one session.
-
-Do **not** create a skill for a one-off task, for something the tools already do
-directly, or for general knowledge the model already has.
-
-## The one rule that matters most
-
-`description` is the only text a future model sees before deciding whether to
-load your skill. Everything else is invisible until then. So it must answer two
-questions in one or two sentences:
-
-1. **What does this do?**
-2. **When should it be used?** — in the words a user would actually type.
-
-Compare:
-
-- Weak: `Helps with PDFs.`
-- Strong: `Fill, flatten and validate AcroForm PDFs. Use when the task mentions
-  PDF forms, field filling, or flattening a fillable PDF.`
-
-The second one will be picked up on a task that says "flatten this fillable
-invoice". The first one will not.
-
-## Writing the body
-
-- Use imperative steps, not background prose. The reader is an agent about to
-  act, not a student.
-- Keep the body under about 5 000 characters. It is loaded whole, every time the
-  skill is used, so length is a recurring cost paid by every future run.
-- Push long reference material into `reference/*.md` and executable steps into
-  `scripts/*`. Those are NOT preloaded — the model reads them on demand with
-  `read_file` or runs them with `bash`. That is the whole point of the three
-  levels, and it is what makes fifty installed skills cost a few thousand
-  characters instead of hundreds of thousands.
-- Say what "done" looks like, and what to do when a step fails.
-
-## Things that do not belong in a skill
-
-- Hostnames, IP addresses, repository URLs, absolute paths from your machine.
-- API keys, tokens, or anything else secret. Skills are plain files that get
-  copied between machines and committed to repositories.
-- Anything that only made sense in the conversation you are in right now.
-
-## Steps
-
-1. Confirm the scope with the user in one sentence: what the skill does and when
-   it fires.
-2. Draft the `description` first and read it back critically. Would a model
-   with only that line pick this skill for the task you have in mind?
-3. Write the body as numbered imperative steps.
-4. Split anything long or reusable into `reference/` and `scripts/` entries and
-   pass them as `files`.
-5. Call `skill_create` with `name`, `description`, `body`, and any `files`.
-6. Immediately call `skill(name="<your-new-skill>")` to read back what you
-   actually wrote. This catches truncated bodies and mangled frontmatter while
-   you can still fix them.
-
-## Naming
-
-Lowercase kebab-case, at most 64 characters: `pdf-forms`, `release-notes`,
-`db-migration-review`. The name becomes a slash command (`/pdf-forms`), so it
-should read as a thing, not as a sentence.
-
-If the name collides with a built-in command, the built-in wins and your skill
-is reachable as `/skill:<name>` instead — so avoid `help`, `model`, `settings`,
-`tools`, `clear`, `reset`, `cwd`, `save`, `resume`, `copy`, `theme`, `thinking`,
-`expand`, `skills`, `exit` and `quit`.
+---
+name: skill-creator
+description: Author a new AragonAgent skill. Use when the user asks to save a procedure as a skill, or when you notice you have re-derived the same multi-step process twice in a session.
+version: 1.1.0
+license: MIT
+activation: auto
+keywords: [meta, authoring, skills]
+# Enforced as a per-turn tool ceiling. `edit_file` is here because revising an
+# existing skill is the single most common thing this skill is asked to do —
+# without it, "fix the description on my deploy skill" fails on the first edit.
+allowed-tools:
+  - read_file
+  - write_file
+  - edit_file
+  - list_dir
+  - skill_create
+---
+
+# Skill Creator
+
+Write a skill that another model instance will actually reach for at the right
+moment. Most bad skills are not badly written; they are badly *described* and
+badly *scoped*.
+
+## When to use
+
+- The user says "remember how to do this", "save this as a skill", or similar.
+- You have worked out a non-obvious, repeatable procedure and expect to need it
+  again.
+- You notice you have rediscovered the same steps twice in one session.
+
+Do **not** create a skill for a one-off task, for something the tools already do
+directly, or for general knowledge the model already has.
+
+## The one rule that matters most
+
+`description` is the only text a future model sees before deciding whether to
+load your skill. Everything else is invisible until then. So it must answer two
+questions in one or two sentences:
+
+1. **What does this do?**
+2. **When should it be used?** — in the words a user would actually type.
+
+Compare:
+
+- Weak: `Helps with PDFs.`
+- Strong: `Fill, flatten and validate AcroForm PDFs. Use when the task mentions
+  PDF forms, field filling, or flattening a fillable PDF.`
+
+The second one will be picked up on a task that says "flatten this fillable
+invoice". The first one will not.
+
+## Writing the body
+
+- Use imperative steps, not background prose. The reader is an agent about to
+  act, not a student.
+- Keep the body under about 5 000 characters. It is loaded whole, every time the
+  skill is used, so length is a recurring cost paid by every future run.
+- Push long reference material into `reference/*.md` and executable steps into
+  `scripts/*`. Those are NOT preloaded — the model reads them on demand with
+  `read_file` or runs them with `bash`. That is the whole point of the three
+  levels, and it is what makes fifty installed skills cost a few thousand
+  characters instead of hundreds of thousands.
+- Say what "done" looks like, and what to do when a step fails.
+
+## Things that do not belong in a skill
+
+- Hostnames, IP addresses, repository URLs, absolute paths from your machine.
+- API keys, tokens, or anything else secret. Skills are plain files that get
+  copied between machines and committed to repositories.
+- Anything that only made sense in the conversation you are in right now.
+
+## Steps
+
+1. Confirm the scope with the user in one sentence: what the skill does and when
+   it fires.
+2. Draft the `description` first and read it back critically. Would a model
+   with only that line pick this skill for the task you have in mind?
+3. Write the body as numbered imperative steps.
+4. Split anything long or reusable into `reference/` and `scripts/` entries and
+   pass them as `files`.
+5. Call `skill_create` with `name`, `description`, `body`, and any `files`.
+6. Immediately call `skill(name="<your-new-skill>")` to read back what you
+   actually wrote. This catches truncated bodies and mangled frontmatter while
+   you can still fix them.
+
+## Naming
+
+Lowercase kebab-case, at most 64 characters: `pdf-forms`, `release-notes`,
+`db-migration-review`. The name becomes a slash command (`/pdf-forms`), so it
+should read as a thing, not as a sentence.
+
+If the name collides with a built-in command, the built-in wins and your skill
+is reachable as `/skill:<name>` instead — so avoid `help`, `model`, `settings`,
+`tools`, `clear`, `reset`, `cwd`, `save`, `resume`, `copy`, `theme`, `thinking`,
+`expand`, `skills`, `exit` and `quit`.