Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    market intelligence curator

    1

    Deduplicate and score corporate signals from RSS feeds for M&A, regulatory, and patent intelligence.

    Free

    Security scanned
    market intelligence curator

    market intelligence curator

    Example session with this skill installed

    Analyze these feed items for 'TechCorp' (ticker: TCORP) and its subsidiary 'CloudScale'. Set minimum_materiality_score to 70 and check for M&A or patent signals.

    • Read your context and instructions
    • Compiled the market intelligence curator

    I have processed the market intelligence surveillance for TechCorp and CloudScale. I identified three unique items, consolidated one duplicate group, and surfaced one high-priority signal.

    Corporate Signal Surveillance Report

    Signal TypeTarget EntityScoreSummary
    M&A_DIVESTITURETechCorp (Parent)90TechCorp announces $2.4B acquisition of AI-Logistics Ltd.
    PATENT_AND_IPCloudScale (Sub)42CloudScale files new patent application for distributed cooling.
    REGULATORYTechCorp (Parent)52FTC requests additional data regarding Q3 filings.

    Duplicate Analysis

    • Canonical Item: "TechCorp to acquire AI-Logistics for $2.4 Billion" (Source: Feed A)
    • Duplicate Found: "TCORP confirms $2.4B AI-Logistics merger" (Source: Feed B)
    • Method: Token 3-gram Jaccard Similarity (Score: 0.88)

    Next steps

    • Review the $2.4B M&A signal as it exceeded the materiality threshold of 70.
    • Adjust minimum_materiality_score to 40 if you wish to capture the lower-scored patent and regulatory items.
    • Provide additional feed_urls for broader surveillance of TechCorp's competitors.

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

    About this skill

    Identify high-value corporate events without manual feed monitoring. This skill processes raw RSS/Atom feeds to detect and score specific signals like M&A activity, regulatory hurdles, and executive turnover. It handles the heavy lifting of deduplication and keyword-based materiality scoring for your target watchlists.

    What it does

    • Multi-source deduplication identifies canonical items using Jaccard similarity and TF-IDF cosine similarity across multiple feeds.
    • Signal scoring applies weighted heuristics to identify M&A (40), regulatory actions (35), patent disputes (35), and leadership changes (25).
    • Entity matching tracks parent companies, subsidiaries, and ticker symbols using strict token boundaries.
    • Materiality filtering excludes low-noise items based on a configurable scoring threshold (default 65).
    • Smart polling utilizes conditional HTTP caching (ETags/If-Modified-Since) to minimize bandwidth and stay within rate limits.

    How it works

    1. Define target profiles by providing parent company names, aliases, and ticker symbols.
    2. Input feed data by supplying parsed feed items or a list of public HTTPS RSS/Atom URLs.
    3. Configure thresholds to set the minimum materiality score for signal detection.
    4. Retrieve structured results including canonical items, duplicate groups, and scored signals.

    Frameworks & tools

    Built for MCP-compliant environments. Uses HTTP caching standards and standard XML feed formats (RSS 2.0, Atom).

    Why this beats prompting it yourself

    Standard LLM prompting often struggles with precise deduplication across hundreds of news items and tends to hallucinate "material" events. This skill uses deterministic similarity algorithms and weighted scoring rules to ensure consistency and prevent token-waste on redundant data.

    Use cases

    • Monitoring M&A activity across a specific industry sector.
    • Tracking regulatory and antitrust developments for a legal watchlist.
    • Surfacing executive turnover signals for competitive intelligence.

    Known limitations

    Does not provide legal or investment advice. Requires public HTTPS URLs for polling and does not support private DNS or redirects. Signals are heuristic candidates, not verified facts.

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

    Requires UTC timestamps ending in Z. Only supports public HTTPS feeds without credentials. Heuristic scores are not legal or investment advice. Does not follow feed redirects.

    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

    Free to use

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 4 days ago

    • Free with an Agensi account

    Permissions

    Terminal / Shell
    Read Files
    Listed4 days ago

    What's inside

    Frequently Asked Questions

    This skill uses Jaccard and TF-IDF similarity algorithms to compare incoming feed items, ensuring that the same news story from multiple outlets is grouped into a single canonical entry rather than flooding your dashboard.

    The curator is built for MCP-compliant environments, making it compatible with any AI agent or platform that supports the Model Context Protocol (MCP) standard for tool integration.

    You define target profiles with company names and tickers, provide the RSS/Atom URLs, and set your materiality threshold; the skill then handles the polling, scoring, and deduplication automatically.

    Unlike a basic prompt that might hallucinate significance, this skill uses deterministic weighted heuristics to score events like M&A or patent disputes, ensuring you only spend tokens on truly material updates.

    The skill currently requires public HTTPS URLs for feed polling and does not support feeds behind private DNS or those requiring complex redirects.

    New in LLM Ops

    Free