agent guild

    1

    Local-first shared memory and task handoff protocol for persistent agent work cycles and session isolation.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    agent guild

    Example session with this skill installed

    I need to step away. Can you save my progress on the React migration task and tell me what the next step is?

    • Read your context and instructions
    • Compiled the agent guild
    • Generated the document

    I have created a checkpoint for your current task. The state, including the completed component refactors and the identified dependency blocks, is now saved to the local guild storage.

    Task Checkpoint: React Migration

    FieldDetails
    Task ID20231027-a1b2c3d4e5f6
    Statusactive
    Projectfrontend-refactor
    Revision5
    SummaryCompleted Button and Input components; AuthContext is pending.

    Progress Report

    • Done: Converted src/components/ui to TypeScript. Verified styles in Storybook.
    • Remaining: Refactor src/context/AuthContext.js to use the new hook pattern.
    • Decisions: Replaced styled-components with Tailwind to match the new design system per your instruction.
    • Evidence: npm test passed for all modified components. See snapshots/rev_5/test-results.log.

    Next steps

    1. Run ag task resume --id 20231027-a1b2c3d4e5f6 when you return to reload this context.
    2. Begin refactoring AuthContext.js following the pattern in src/hooks/useAuth.ts.
    3. Update the provider in App.tsx once the context is stable.

    App.tsx

    PDF · document

    Generated

    Example file from a real run - the skill writes it into your workspace.

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

    About this skill

    Context window limits and session resets often cause agents to lose track of multi-step work. You end up repeating instructions or manually copying progress between chats.

    What it does

    • Persistent memory manages long-term facts, identities, and rules across different sessions and agents.
    • Task checkpointing saves the full state of active work including goals, decisions, and evidence into isolated task packages.
    • Opt-in resumption prevents context pollution by listing pending tasks and only loading one when explicitly selected.
    • Cross-agent messaging enables structured communication between different agent identities without duplicating entire contexts.
    • History management enforces a configurable storage budget (default 1GiB) with preview-first maintenance plans to prevent bloat.

    How it works

    1. Bootstrap identity by loading agent-specific rules and universal preferences without injecting old task data.
    2. List pending work using metadata-only queries to show active tasks and their status for user selection.
    3. Resume task context to pull in the specific history, attachments, and next actions for the chosen work item.
    4. Checkpoint progress during the session to save milestones, changed files, and updated implementation details.
    5. Archive on completion to move the entire task package to a long-term storage directory for audit and reference.

    Frameworks & tools

    The skill is built with a Python 3.9+ CLI and utilizes standard library modules. It operates entirely on local filesystem storage within ~/.agent-guild/ or a custom directory.

    Why this beats prompting it yourself

    Manual prompts for "remembering" things are fragile and often get truncated or ignored as conversations grow. This skill provides a structured CLI-backed protocol that ensures work state is saved to disk, not just held in volatile model context.

    Use cases

    • Manage long-running software migrations across multiple days and sessions.
    • Handoff complex research tasks from one specialized agent to another.
    • Maintain a consistent persona and set of coding standards across all local projects.
    • Audit past decisions by reviewing archived task packages and implementation evidence.

    Known limitations

    Does not provide distributed locking for multi-device sync or automated credential management. Requires Python 3.9+ available on the host machine.

    Requires Python 3.9+ on the host. No distributed-lock for multi-device sync. Does not manage credentials or provide OS-level access controls.

    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

    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 29 days ago

    Requires

    GitHub
    Skillhub
    Clawhub
    Python
    macOS
    Git

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Browser
    Network Access
    Environment Variables

    Allowed Hosts

    raw.githubusercontent.com
    api.skillhub.cn
    github.com
    clawhub.ai
    Listed29 days ago

    What's inside

    Frequently Asked Questions

    Agent Guild functions as a local protocol that manages task state and long-term memory outside of the model's context window. It prevents "context drift" by allowing agents to save their progress to disk and resume specific tasks with all relevant history and rules intact.

    This skill is designed to be framework-agnostic and works with any AI agent capable of executing Python 3.9+ CLI commands. Because it relies on a local filesystem protocol rather than specific API integrations, you can use it to coordinate work between different models or platforms.

    All data, including task checkpoints and agent identities, is stored locally on your machine (defaulting to `~/.agent-guild/`). It does not require a cloud subscription or external database, ensuring your work cycles remain private and under your control.

    Installation is straightforward via a Python-based setup, and the package includes the core CLI tool, the handoff protocol documentation, and pre-configured templates for agent identities. Once installed, you can immediately begin checkpointing tasks and managing persistent memory.

    Unlike simple prompts that get buried as a conversation grows, this skill provides a structured storage budget and "preview-first" maintenance. It allows you to load only the specific task data you need, preventing the performance degradation and "hallucinations" caused by overloaded context windows.

    New in Productivity