planning with files

    1

    Externalize agent memory to Markdown files to manage complex, multi-step tasks without losing context or focus.

    $1.99

    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

    planning with files

    Example session with this skill installed

    I need to migrate our auth logic from JWT to OAuth2. It involves 4 files and new environment variables. Start a session using the planning-with-files protocol.

    • Read your context and instructions
    • Compiled the planning with files

    I have initialized the planning environment in your project root. I created task_plan.md, findings.md, and progress.md based on the Manus-style templates.

    Initialized Planning Files

    FileStatusInitial Content
    task_plan.mdACTIVEMigration phases: Discovery, Env Setup, Implementation, Testing.
    findings.mdOPENAwaiting OAuth2 provider specs and redirect URI requirements.
    progress.mdOPENSession started. Current Directory: /project/root.
    
    # Task Plan: JWT to OAuth2 Migration
    
    ## Goal
    Replace custom JWT authentication with a standardized OAuth2 flow across the API and Frontend.
    
    ## Phases
    1. [in_progress] Discovery: Map existing auth endpoints and identify OAuth2 provider (Google/GitHub).
    2. [todo] Environment Setup: Configure client IDs, secrets, and callback URLs.
    3. [todo] Implementation: Update `auth.middleware.ts` and login routes.
    4. [todo] Testing: Verify token exchange and session persistence.
    
    ## Errors Encountered
    
    | Error | Attempt | Resolution |
    | :--- | :--- | :--- |
    | N/A | 1 | N/A |
    
    

    Next steps

    • I will now use grep to locate all instances of JWT handling in the codebase.
    • I will record the specific files needing modification in findings.md.
    • I will search for the OAuth2 provider documentation to confirm the necessary library updates.

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

    About this skill

    Large language models lose focus during long-running tasks as the context window fills with tool outputs and logs. This skill implements a file-based state management system that uses your local filesystem as external RAM. It forces the agent to document plans, research, and progress in persistent Markdown files, ensuring it never loses track of the original goal or repeats failed attempts.

    What it does

    • Persistent Planning creates and maintains a task_plan.md to track phases, goals, and completion status.
    • Knowledge Retention saves research and technical discoveries to findings.md to prevent multimodal data loss.
    • Session Logging records tool outputs and test results in progress.md for a complete audit trail.
    • Error Tracking maintains a structured log of failures and resolutions to avoid repeating unsuccessful actions.
    • State Recovery enables the agent to resume complex workflows by reading plan files after context resets.

    How it works

    1. Initialize by creating the three core planning files in the project root using provided templates.
    2. Execute tasks while following the 2-action rule, writing browser or search findings to disk every two operations.
    3. Monitor progress by updating phase statuses and logging any errors encountered during tool use.
    4. Iterate using a 3-strike protocol that forces a change in approach after failures rather than silent retries.

    Frameworks & tools

    This skill uses standard Markdown for state storage and relies on Bash, Read, Write, and Edit tools. It includes helper scripts for session initialization and completion verification.

    Why this beats prompting it yourself

    Generic prompts often fail to maintain long-term organization as the conversation grows. This skill installs hard constraints and hooks that trigger file updates automatically, ensuring the agent uses the filesystem for memory instead of relying on a volatile context window.

    Use cases

    • Multi-step Refactoring where changing one component impacts several others across a codebase.
    • Deep Research involving multiple web searches and documentation lookups that need synthesis.
    • Complex Debugging requiring tracking of multiple hypotheses and attempted fixes.

    Known limitations

    Requires write access to the local filesystem. Planning files must be manually created in the project root to function correctly.

    Requires filesystem write access. Planning files must be in the project root. Optimal for tasks requiring more than 5 tool calls.

    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 8 days ago

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Network Access
    Listed8 days ago

    What's inside

    Frequently Asked Questions

    Externalize agent memory to Markdown files to manage complex, multi-step tasks without losing context or focus.

    This skill uses the open SKILL.md standard. Use it through an MCP-compatible agent connected to Agensi, and check the compatibility information on this page for confirmed agents and required capabilities.

    Connect your AI agent to Agensi through MCP, then ask it to use planning with files on your task. Your agent loads the latest approved version each time. Paid skills require a purchase or active subscription. Skill files are not available for download; your licence does not allow you to save, copy or export any part of a skill.

    New in Agents & Orchestration