7.9
/ 10
1 evaluations
1.8k Downloads
Overview
Provide a robust integration with the Fabric HTTP API to create, search, and manage Fabric workspace resources (notepads/notes, folders, bookmarks, files, tags) from within OpenClaw agents.
Key Advantages
1.Direct, first-class integration with Fabric’s official HTTP API, covering core entities like notepads, folders, bookmarks, and tags.
2.Clear, opinionated guidance on common pitfalls (e.g., using POST /v2/notepads instead of non-existent /v2/notes, tag object shapes, alias vs UUID parent IDs).
3.Cross-platform helper scripts for Node and Python, designed to avoid bash-only wrappers and work reliably on most systems.
4.Well-documented default behaviors (e.g., mapping user ‘title’ to API field ‘name’, defaulting parentId to @alias::inbox when unspecified).
5.Concrete examples for core workflows (create note, folder, bookmark; filter resources; search), lowering the chance of schema/validation errors at runtime.
","Explicit error-handling guidance (400 vs
Use Cases
- Creating new notes/notepads in a user’s Fabric inbox or specified folders based on conversational input or task outputs.
- Organizing a user’s Fabric workspace by creating and nesting folders, then moving or creating resources under those folders.
- Capturing URLs from a conversation and saving them as bookmarks with appropriate tags and metadata in Fabric.
- Browsing and listing resources in specific folders (by UUID) to summarize, audit, or present organized views of a user’s content.
- Semantic or text-based search over Fabric notes (e.g., “find the meeting notes about project X”) and then acting on the results (summarize, link, update).","Automated knowledge management routines, eg
Evaluation Scores
7.9
/ 10
Reliability
7.5
Functionality
8.5
Usability
8.0
Safety
7.5
Performance
8.0
Compatibility
8.0
Based on 1 evaluation · Latest: 3/20/2026
Download Trend
Loading...
Evaluation History (1)
7.9/103/20/2026▼
OS: darwin-x64LLM: deepseek/deepseek-v3.2
**Judgement:** This skill is a solid, well-documented integration for working with a user’s Fabric workspace via the official HTTP API. It’s particularly suitable when an agent needs persistent, structured knowledge storage and retrieval in Fabric (notes, folders, bookmarks, and search) rather than ephemeral, in-memory context.
**Strengths:**
- Good coverage of key Fabric resources (notepads/notes, folders, bookmarks, tags) with clear examples for each core workflow.
- Cross-platform Node and Python helpers make HTTP usage straightforward in typical OpenClaw environments.
- Documentation explicitly calls out critical schema gotchas (e.g., `name` vs `title`, tag object shape, alias vs UUID for `parentId`), which significantly reduces runtime errors.
- Provides practical guidance on error handling and retries, especially around 400, 429, and 5xx responses.
**Risks / Limitations:**
- Requires a valid `FABRIC_API_KEY`; misconfiguration will lead to 401/403 and non-functional behavior until fixed.
- Operations are inherently stateful and potentially destructive: incorrect IDs, folders, or tags could clutter or misstructure the user’s Fabric workspace; care is needed with any future delete/update calls.
- The `/v2/resources/filter` endpoint requires UUID `parentId` rather than alias, which is an easy source of subtle bugs if the agent ignores the docs.
- Reliability is bounded by Fabric’s own uptime, rate limiting (429s), and the quality of the OpenAPI spec; heavy workloads may hit rate limits.
**Recommended Scenarios:**
- Agents that act as a “knowledge worker” for the user: capturing conversation outputs, summaries, and links into Fabric as persistent notes or bookmarks.
- Personal organization or PKM agents that create and manage folder hierarchies and keep Fabric content sorted.
- Retrieval-augmented workflows that need to search existing Fabric notes/bookmarks by text and then present or transform those results.
- Any environment where the user is already committed to Fabric as their primary notes/bookmarks hub and wants the agent to operate directly on that system.
Comments (0)
No comments yet. Be the first!