Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Persona Flow AI

    1

    A discovery agent that recommends activities from a JSON catalog using strict budget and preference matching logic.

    Free

    1 useSecurity scanned
    Persona Flow AI

    Persona Flow AI

    Example session with this skill installed

    I love art and nature, my budget is 45,000 IDR, I have 90 minutes, and I want something relaxed.

    • Read your context and instructions
    • Compiled the persona flow ai

    I have analyzed your preferences for art and nature with a 45,000 IDR budget and a relaxed mood. Based on the catalog, here are two activities that fit your criteria:

    Recommended Activities

    IDActivity NameEstimated Cost (IDR)DurationMatch Reason
    ACT-003Beginner Photography Walk0 - 40,00090 minsFits your art/nature interests and relaxed mood within your 45k budget.
    ACT-005Nature Sketching10,000 - 45,00075 minsCombines art and nature in a quiet, relaxed setting at your exact budget limit.

    Next steps

    • Let me know if you would like to see more details on the Beginner Photography Walk.
    • If you have more time available, I can search for longer workshops in the medium budget range.
    • Tell me if you want to switch to a more social or energetic atmosphere.

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

    About this skill

    Personalized recommendations often suffer from hallucinations or budget creep. This skill enforces strict grounding by matching user preferences against a specific activity catalog while adhering to hard budget constraints.

    What it does

    • Preference Discovery gathers interests, maximum budget, duration, and desired mood through natural dialogue.
    • Strict Budgeting treats price limits as hard constraints, filtering out any activities where the maximum cost exceeds the user's cap.
    • Catalog Grounding ensures every recommendation exists in the provided reference data, preventing fabricated activity details.
    • Contextual Matching prioritizes interest and atmosphere to explain exactly why a specific recommendation fits the user's profile.
    • Constraint Handling identifies when no matches exist and prompts the user to adjust specific preferences rather than guessing.

    How it works

    1. Initiate a conversation to identify the user's primary activity interests.
    2. Collect remaining constraints including maximum IDR budget, available time in minutes, and preferred mood.
    3. Query the persona_catalog.json to identify activities that fit all four criteria simultaneously.
    4. Present up to three verified recommendations with their specific catalog IDs and cost ranges.

    Frameworks & tools

    The skill is designed for agents with access to JSON-based reference data. It uses standard logic for filtering and matching against interests, budget_level, duration_minutes, and moods fields.

    Why this beats prompting it yourself

    Generic prompts often ignore "max budget" limits or invent activities that sound plausible but do not exist. This skill uses a logic-heavy approach to ensure 100% catalog integrity and strictly enforces budget caps to prevent low-value suggestions.

    Use cases

    • Building a demonstration for a personalized travel or event discovery app.
    • Testing RAG-style constraint satisfaction in a conversational interface.
    • Implementing a controlled recommendation engine for a closed set of partner activities.

    Known limitations

    This skill only recommends activities from the provided fictional catalog and cannot access external real-world event databases. It requires all four user preferences to be defined before suggesting matches.

    Requires fictional persona_catalog.json for operation. Cannot access real-world event APIs. Will not recommend activities until all four preference fields are populated or explicitly waived.

    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 any payment.

    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

    1 use

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 26 days ago

    • Free with an Agensi account

    Permissions

    Read Files

    File Scopes

    gemini/skill.md
    Listed26 days ago

    What's inside

    Frequently Asked Questions

    Unlike generic prompts that might suggest expensive or non-existent options, this skill uses strict logic to ensure every recommendation is both within your specified budget and verified against your JSON catalog.

    This skill is designed for any AI agent framework capable of processing and querying structured JSON data, making it ideal for developers building RAG-style conversational interfaces.

    The package includes the core logic for the recommendation engine, the 'persona_catalog.json' template, and a set of instructions for integrating the discovery dialogue into your agent.

    Yes, the skill is built to be flexible; you can easily swap the provided sample activities for your own partner data, provided you maintain the standard JSON fields like price, duration, and mood.

    The agent is programmed to handle 'no-match' scenarios gracefully by identifying which specific constraint—such as budget or time—is blocking the results and asking the user to adjust it.

    No, this skill is a 'closed-loop' discovery agent designed for high integrity within a specific dataset; it does not query live external APIs or web search results.

    New in Agents & Orchestration