Multi-user personal knowledge repository with hybrid search (SQLite FTS5 + Gemini embeddings), RSS feed integration, AI-powered summarization, and MCP server for programmatic access to bookmarks and repository queries.
OmniLink AI provides 8 well-named tools with clear verb-noun patterns (search_repository, save_bookmark, etc.). All tools have descriptions (avg ~120 chars, within baseline 34-392 range). Input schemas are present with types and descriptions for most parameters. Parameter descriptions are adequate but lack actionable constraints (e.g., no format hints for URLs, no enum validation for category/platform beyond search tool). Error handling is minimal, no recovery guidance or actionable error messages visible in source. Tool composition is sound (single responsibility), but descriptions could be more LLM-optimized (e.g., 'ask_repository' lacks clarity on when to use vs search_repository).
Ask a natural language question about your bookmarks and repository. The AI will search relevant bookmarks and synthesize an answer with citations.
Delete a bookmark from the repository by ID.
Retrieve a specific bookmark by ID with full metadata, summary, and reading status.
Get aggregate statistics about your repository including total bookmarks, categories, platforms, and reading progress.
List all bookmarks in the repository with optional filtering by category, platform, or read status.
Save a new URL or repository into OmniLink with automated AI categorization, summary extraction, and vector indexing.
Output schemas not documented. LLMs cannot plan downstream calls or extract required fields (e.g., what does search_repository return? Does it include bookmark IDs for chaining to get_bookmark?).
Error handling absent from visible code. No recovery guidance (e.g., 'If bookmark not found, try search_repository()'), no error classification (retryable vs fatal), no actionable error messages.
ask_repository tool lacks clarity on differentiation from search_repository. Both appear to query bookmarks; descriptions do not explain when to use each.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 54 | 2026-07-28+ | v2 |
Search personal bookmarks and repositories using Hybrid Search (SQLite FTS5 BM25 + Gemini Dense Vector Embeddings + Reciprocal Rank Fusion).
Update an existing bookmark's metadata including title, notes, tags, category, and read status.
Parameter descriptions lack actionable constraints. E.g., 'category' in save_bookmark is optional but no enum or valid values listed; 'platform' in search_repository has enum but save_bookmark does not.
Destructive tool (delete_bookmark) lacks confirmation/dry-run pattern.