AI Agent Hub
Back to plugins
⚙️

task-passport

Workflow Updated 2026.08.27

Run the following command in DeepSeek Harness:

dsh plugin install dongsheng123132/task-passport

Paste the following prompt into your AI chat to install this plugin:

Run dsh plugin install dongsheng123132/task-passport in your DeepSeek Harness terminal to complete the installation, and visit https://github.com/dongsheng123132/task-passport for the complete source code.

About this plugin

In multi-AI workflows, developers frequently hit a wall when handing off tasks: the only reliable method has been copying entire chat histories. This approach buries critical context under irrelevant chatter, overwhelms new models with noise, and often leads to fragmented progress. Task Passport changes this paradigm. Instead of shipping conversations, it passes along the precise “current state of the world” for any given objective, ensuring continuity without the baggage.

At its core, the plugin assigns every ongoing goal a persistent “Task Passport”—a versioned, secure record holding verified facts, reasoning, targets, and next steps. Through a straightforward CLI, users can query status, save milestones, and export work into portable single-file bundles called TaskPacks for cross-machine or cross-person handovers. Built on strict safety principles, packs automatically downgrade unverified machine facts at the source, refuse to ship unresolved prompts, and contain pure data rather than executable instructions. State lives independently in your preferred provider, making it a lightweight, always-available bridge across sessions.

Whether you’re orchestrating complex projects across Claude Code, Codex, and DeepSeek Harness, or simply want to stop restarting from scratch after a session ends, Task Passport is your workflow anchor. It’s particularly valuable for developers managing long-form documents, iterating through multi-agent reviews, or coordinating human-in-the-loop sign-offs. By decoupling task intent from ephemeral chats, it turns chaotic handoffs into predictable, trackable transitions—so you can focus on building, not rebuilding.

Use Cases

  • Seamlessly transfer code progress and pending tasks when switching between different AI coding assistants.
  • Package intermediate states of long-term documentation or research projects for colleagues to continue work.
  • Log human feedback in human-in-the-loop workflows and send structured replies directly back to AI agents.

Best For

  • Developers managing complex projects that require frequent handoffs between multiple AI agents.
  • Engineers seeking to eliminate chat log clutter and pursue clean, efficient debugging experiences.
  • Agile technical teams engaged in long-cycle development or requiring cross-team progress sharing.