Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Session Handoff Lite

    1

    One session ends as the next one starts working immediately

    Free

    1 useSecurity scanned
    Session Handoff Lite

    Session Handoff Lite

    Example session with this skill installed

    I'm hitting the context limit. Wrap this session up into a handoff. We were halfway through migrating the auth logic to the new provider.

    • Read your context and instructions
    • Compiled the session handoff lite

    I have generated the HANDOFF.md file at the repository root to capture the current progress of the authentication provider migration.

    HANDOFF.md

    1. Objective and state

    Goal: Migrate authentication logic from LegacyAuth to NewProvider.
    Status: In progress. The core provider interface is implemented in src/auth/new_provider.ts, but call sites in the middleware are still using the legacy implementation.

    2. Decisions and why

    Chose manual migration over automated codemods because the legacy signatures vary significantly between the API and the internal dashboard.
    Chose to keep both providers active during the transition to allow for a phased rollout and easier rollbacks.

    3. The "do not redo" list

    Do not attempt to use provider-utils.v2.js for this migration. I tried integrating it for 30 minutes, but it has a hard dependency on Node 14 globals that break our current CI pipeline.

    Do not refactor the User interface yet; changing the ID type now causes a cascade of failures in the reporting service.

    4. Next actions with acceptance criteria

    1. Migrate middleware/auth.ts: Update the verifySession function to use NewProvider.validate().
      • Acceptance: npm run test:unit middleware passes.
    2. Update routes/login.ts: Replace LegacyAuth.login() with NewProvider.authenticate().
      • Acceptance: Manual login via /api/login returns a JWT with the new header format.

    5. Verification recipe

    Run the following to verify the current broken state:

    npm run test:auth
    

    Expected result: 12 tests passing, 4 failing (specifically the middleware integration tests). If more than 4 fail, check if the .env.local provider keys have expired.

    Next steps

    1. Commit the HANDOFF.md file to the branch.
    2. Start the next session by feeding this file to the agent.
    3. Begin with the middleware migration as outlined in the next actions.
    HANDOFF.md

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

    About this skill

    session-handoff-comparison.png

    The problem. Context fills up mid-refactor, or you stop mid-debug, and the next session spends its first couple of turns re-reading code to reconstruct state, re-trying an approach that already failed, or "fixing" something that was chosen the other way on purpose. What gets lost is never the obvious stuff, it's the decisions and the dead ends.

    What it does. When a session is ending, this skill writes a single HANDOFF.md with five sections: objective and state, decisions and why, the do-not-redo list, next actions with acceptance criteria, and a verification recipe. The bar it holds itself to: a fresh session reads it and starts working without asking a single question.

    Why this beats prompting it yourself. Ask an agent to "summarize where we are" and you get a narrative with the uncertainty baked in. The failure modes are specific: dead ends vanish, decisions lose their rationale, state becomes vague. This skill inverts the priorities, failed attempts are treated as the most valuable content, decisions carry their reasons, and state is pinned to SHAs, paths and commands.

    What's inside

    • SKILL.md — the five-section method, the writing rules, and a worked example.

    limitations

    • Lite edition: covers a single session's wrap-up. Environment capture, open-questions handling, a fill-in template, and worked multi-agent scenarios are in the full Session Handoff skill.
    • It writes the document; it doesn't archive your previous handoffs for you.
    • It cannot recover context that was already lost before it ran, run it before compaction, not after.

    FAQ

    Is this really free? Yes. It is a deliberate cut of the full Session Handoff skill: same method, one session, one file, no template or scripts.

    What's the difference from the full Session Handoff skill? The full edition adds environment capture (processes, ports, env vars), open-questions handling, a fill-in template, worked scenarios including multi-agent pickup, and a handoff history policy.

    Does this work in Cursor, Codex or Copilot? Yes — plain instructions with no runtime; it works anywhere the skill format is supported.

    Where does the handoff file go? HANDOFF.md at the repository root, committed with the work it describes.

    What if the session achieved nothing? Then the handoff is short: objective, blocked status, and the do-not-redo list — the valuable output of a session that failed.

    Requires local file system access to write the HANDOFF.md file. Does not include environment variable capture or complex multi-agent orchestration.

    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 1 month ago

    • Free with an Agensi account

    Permissions

    Write Files
    Listed1 month ago

    What's inside

    Frequently Asked Questions

    This skill acts as a technical bridge between sessions by generating a structured HANDOFF.md file that documents work state, failed attempts, and specific next steps to eliminate the "re-learning" phase when starting fresh.

    Yes, this skill is language-agnostic and designed to work in any CLI or IDE environment, provided the agent has access to your local file system to save the markdown file.

    Unlike generic prompts, this skill enforces a strict schema that includes 'Dead-end documentation' and 'Verification recipes,' ensuring the output is technically actionable rather than just a narrative summary.

    The Lite version focuses on manual triggers and core state tracking, whereas the full version includes automated environment capture and advanced multi-agent delegation logic.

    The skill captures recently changed files, commit SHAs, and terminal outputs from your current session to ensure the next agent knows exactly where the last one left off.

    New in Productivity