Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    obsidian notes agent

    1

    Automated Obsidian note creation and hybrid search with built-in hallucination checks and prompt injection defense.

    $9.99

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    obsidian notes agent

    obsidian notes agent

    Example session with this skill installed

    Search my vault for notes about "vector databases" and summarize what I've collected so far. Also, check if there are any action items from my last meeting on this topic.

    • Read your context and instructions
    • Compiled the obsidian notes agent

    I searched your Obsidian vault using the hybrid BM25+RRF retrieval system. I found 3 relevant notes and identified one pending action item from your recent meeting.

    Knowledge Synthesis: Vector Databases

    Note SourceKey ConceptStatus
    2024-03-10-rag-basics.mdDiscussion on Pinecone vs Milvus for production RAG.Verified
    2024-03-12-indexing-strategies.mdImplementation details for HNSW and IVF-Flat indexing.Verified
    2024-03-15-meeting-architecture.mdDecision to use pgvector for the initial MVP.Needs Review

    Summary of Findings

    Your notes indicate a preference for starting with pgvector due to existing PostgreSQL infrastructure, despite evaluating dedicated solutions like Pinecone. The technical notes emphasize HNSW as the preferred indexing strategy for low-latency retrieval.

    Action Items

    From 2024-03-15-meeting-architecture.md: @DevTeam to benchmark pgvector performance with 1M+ embeddings by next Friday.

    Next steps

    • Run buscar_notas(consulta="pgvector benchmarks") to see if the performance tests started.
    • Create a new MOC (Map of Content) for "Vector Database Research" to link these files.
    • Use resumir_reuniao on the transcript of today's follow-up to update these records.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Managing a massive Obsidian vault manually leads to fragmented ideas and forgotten insights. This skill automates note creation and retrieval using a LangGraph pipeline and MCP server to turn raw text into verified, structured knowledge.

    What it does

    • Note creation sanitizes input, strips prompt injections, and saves formatted Markdown directly to your vault.
    • Hybrid search retrieves notes using BM25 and character n-grams with Reciprocal Rank Fusion for high-relevance results.
    • Fidelity verification flags claims that cannot be supported by the source text for manual review.
    • Injection defense automatically purges malicious override instructions like "ignore previous prompts" from note content.
    • Meeting summarization converts transcripts into action items and decisions with structured frontmatter.

    How it works

    1. Capture: Provide a transcript, article, or raw text and request it be saved to Obsidian.
    2. Process: The agent runs the content through the criar_nota tool to sanitize, verify, and format the data.
    3. Review: Check the status report for flagged unsupported claims or stripped injections.
    4. Retrieve: Query your vault using buscar_notas to find and synthesize existing knowledge for new notes.

    Frameworks & tools

    Built for Obsidian using the Model Context Protocol (MCP). It utilizes a LangGraph pipeline for logic flow and implements BM25 + RRF for advanced search ranking.

    Why this beats prompting it yourself

    Standard prompts often hallucinate facts or fail to maintain consistent frontmatter schemas. This skill uses a specialized MCP server to interface directly with your local file system while enforcing strict fidelity checks and security guardrails that a simple chat interface cannot provide.

    Use cases

    • Research synthesis by saving verified highlights from multiple articles into a connected second brain.
    • Meeting automation to transform raw transcripts into actionable, clean Obsidian notes.
    • Knowledge retrieval when you need to find specific concepts across thousands of legacy notes.
    • Secure note-taking for processing untrusted or third-party content without risking prompt injection.

    Known limitations

    Requires the notes-agent MCP server to be active. It cannot access absolute file system paths or display full paths for security reasons.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Productivity.

    Requires the notes-agent MCP server for vault access. Limited to file system interactions via the provided toolset. Cannot bypass automated injection filtering.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 3 days ago

    Software
    PythonGit

    Permissions

    Read Files
    Network Access
    Listed3 days ago

    What's inside

    Frequently Asked Questions

    This skill automates the creation and retrieval of Obsidian notes while using a LangGraph pipeline to verify that every saved claim is supported by the source text, preventing the common AI issue of 'hallucinated' facts.

    The skill requires an active 'notes-agent' MCP (Model Context Protocol) server to interface with your local files. It is compatible with any AI agent or platform that supports MCP connections and LangGraph workflows.

    Unlike standard prompts, this skill uses a multi-layered security layer that automatically detects and strips malicious override instructions (like 'ignore previous instructions') from raw text before it ever reaches your vault.

    You will receive the pre-configured LangGraph pipeline, the specialized MCP server integration tools, and documentation for connecting the agent to your local Obsidian vault.

    The agent uses Hybrid Search, combining BM25 keyword matching with character n-grams and Reciprocal Rank Fusion (RRF), allowing it to find relevant notes even if you can't remember the exact phrasing or keywords.

    Yes, the agent is specifically designed to process raw transcripts, automatically extracting action items and decisions while formatting them with consistent frontmatter for your Obsidian workspace.

    New in Productivity