@cteno/sdk-linux-x64 0.0.0-stage → 2.10.14

package/tools/image_generation/TOOL.md

@@ -0,0 +1,81 @@
+---
+id: image_generation
+name: Image Generation
+description: Generate images using text-to-image models via Happy Server proxy
+category: system
+version: "2.0.0"
+supports_background: true
+should_defer: true
+search_hint: "generate image text-to-image AI drawing"
+is_read_only: false
+is_concurrency_safe: false
+---
+
+# Image Generation Tool
+
+Generate images from text descriptions. Images are generated via Happy Server proxy and billed per image from user balance.
+
+## Available Models & Pricing
+
+| Model | Price | Description |
+|-------|-------|-------------|
+| `qwen-image-max` (default) | ¥0.20 | 高质量图像 (固定尺寸) |
+| `qwen-image-2.0-pro` | ¥0.20 | 最强文字渲染 + 自定义分辨率 |
+| `qwen-image-plus` | ¥0.08 | 快速图像生成 |
+| `wan2.6-t2i` | ¥0.08 | 写实摄影 + 文字渲染 |
+
+## Parameters
+
+```json
+{
+  "prompt": "string (required) - Image description in Chinese or English",
+  "model": "string (optional) - Model name, default: qwen-image-max",
+  "size": "string (optional) - Image size, default: 1280*1280. Supported: 1280*1280, 1024*1024, 720*1280, 1280*720",
+  "negative_prompt": "string (optional) - What to avoid in the image",
+  "seed": "integer (optional) - Random seed for reproducibility (0-2147483647)",
+  "notify": "boolean (optional) - Send notification when done, default: true",
+  "timeout": "integer (optional) - Timeout in seconds, default: 300"
+}
+```
+
+## Examples
+
+**Basic usage** (returns immediately with run_id):
+```json
+{
+  "prompt": "一只可爱的橙色小猫坐在窗台上，阳光洒在它的毛发上，水彩画风格"
+}
+```
+
+**With custom parameters**:
+```json
+{
+  "prompt": "未来城市的全景图，赛博朋克风格，霓虹灯闪烁，超高清细节",
+  "model": "qwen-image-2.0-pro",
+  "size": "1280*720",
+  "negative_prompt": "模糊，低质量，变形，扭曲",
+  "seed": 42
+}
+```
+
+## Output
+
+Returns run_id immediately, then sends notification with local file path when complete.
+
+## Notes
+
+- **Background Only**: All executions run in background, typically takes 10-60 seconds
+- **Auto Download**: Images are automatically downloaded to the current working directory
+- **Billing**: Charged per image from user balance
+- **Prompt Tips**:
+  - Be specific about style, composition, lighting
+  - Use Chinese for better results
+  - Include quality keywords like "高清", "细节丰富", "专业摄影"
+
+## Agent Usage
+
+When user requests image generation:
+1. Call this tool with the enhanced prompt
+2. The tool generates the image and **automatically downloads it to the current working directory**
+3. You will receive a notification with the local file path when complete
+4. Inform the user about the saved file location

package/tools/skill/TOOL.md

@@ -0,0 +1,117 @@
+---
+id: "skill"
+name: "Skill"
+description: |
+  Skill runtime management: activate/deactivate skills in the current conversation,
+  list installed skills, search/install from SkillHub.
+  To create new skills, activate the 'skill-create' skill instead.
+category: "system"
+version: "2.0.0"
+supports_background: false
+full_disclosure: always
+input_schema:
+  type: object
+  properties:
+    operation:
+      type: string
+      enum: ["list", "activate", "deactivate", "search", "browse", "install"]
+      description: |
+        list: Show all installed skills
+        activate: Load skill instructions into conversation context
+        deactivate: Mark skill as no longer active
+        search: Search SkillHub registry
+        browse: View popular/trending skills from SkillHub
+        install: Download and install from SkillHub
+    id:
+      type: string
+      description: "Skill ID (for activate/deactivate)"
+    skill_id:
+      type: string
+      description: "Alias for id (for activate/deactivate)"
+    query:
+      type: string
+      description: "Search query (for search operation)"
+    slug:
+      type: string
+      description: "SkillHub skill slug (for install operation)"
+    args:
+      type: string
+      description: "For activate: arguments passed to the skill (available as $ARGS/$1 in instructions)"
+    include_resources:
+      type: boolean
+      description: "For activate: include file tree (default: true)"
+    reason:
+      type: string
+      description: "For deactivate: optional reason"
+    limit:
+      type: number
+      description: "Max results for search (default 20)"
+  required:
+    - operation
+is_read_only: false
+is_concurrency_safe: false
+---
+
+# Skill Tool
+
+Runtime tool for skill activation and discovery.
+
+## Quick Reference
+
+| Operation | Required Params | Description |
+|-----------|----------------|-------------|
+| list | - | Show all installed skills with descriptions |
+| activate | id | Load skill instructions into context |
+| deactivate | id | Mark skill as inactive |
+| search | query | Search SkillHub for skills |
+| browse | - | View popular/trending skills |
+| install | slug | Install skill from SkillHub |
+
+## Variable Substitution
+
+When a skill is activated, the following variables in SKILL.md are automatically replaced:
+- `${SKILL_DIR}` → skill directory absolute path
+- `${SESSION_ID}` → current session ID
+- `${SKILL_ID}` → skill ID
+- `${SKILL_NAME}` → skill display name
+- `$ARGS` / `$1` → arguments passed via the `args` parameter
+
+## Script Auto-Execution
+
+Shell commands embedded in SKILL.md are executed at activation time:
+- Fenced blocks: `` ```! command ``` `` → output replaces the block
+- Inline: `` !`command` `` → output replaces inline
+
+## Three-Layer Loading
+
+Skills are loaded from these sources (higher priority overrides lower):
+1. **builtin** — shipped with Cteno
+2. **global** — `~/.agents/skills/` (Agent Skills standard), then `~/.cteno/skills/`
+3. **workspace** — `{workdir}/.agents/skills/` (standard), then `{workdir}/.cteno/skills/` (project-specific, highest)
+
+## Examples
+
+### Activate a skill
+```json
+{"operation": "activate", "id": "xlsx"}
+```
+
+### Activate with arguments
+```json
+{"operation": "activate", "id": "xlsx", "args": "analyze sales data"}
+```
+
+### Search SkillHub
+```json
+{"operation": "search", "query": "pdf generation", "limit": 10}
+```
+
+### Install from SkillHub
+```json
+{"operation": "install", "slug": "pdf-generator"}
+```
+
+### Create new skill
+```json
+{"operation": "create", "name": "my-custom-skill"}
+```

package/tools/update_profile_notes/TOOL.md

@@ -0,0 +1,39 @@
+---
+id: "update_profile_notes"
+name: "Update Profile Notes"
+description: "Update your own behavior notes based on user interactions"
+category: "agent_profile"
+version: "1.0.0"
+supports_background: false
+should_defer: true
+search_hint: "update profile behavior notes communication style"
+input_schema:
+  type: object
+  properties:
+    notes:
+      type: string
+      description: "Updated behavior notes (replaces existing notes entirely)"
+  required:
+    - notes
+is_read_only: false
+is_concurrency_safe: false
+---
+
+# Update Profile Notes Tool
+
+Update your behavior notes based on user feedback, preferences, and interaction patterns.
+
+## When to Use
+
+Use this tool when:
+- The user shares preferences about communication style
+- You learn important traits to maintain consistency
+- The user explicitly asks you to remember behavioral preferences
+
+## Format
+
+Write profile notes as concise bullet points:
+- Communication style preferences
+- Domain expertise areas
+- Decision-making principles
+- Response format preferences