ccch1mneyyy/working-activity

Lively Working-line extension for pi CLI and DSH

What it solves

This project addresses the lack of visibility into an AI agent's internal state during execution. Instead of a generic loading spinner, it provides a dynamic "working status line" that shows exactly what the agent is doing in real-time, including tool usage, progress, and internal reasoning summaries.

How it works

The tool functions as an extension/plugin for two platforms: pi CLI and DeepSeek Harness (DSH). It monitors agent events (such as tool execution starts, updates, and ends) and parses the model's streaming output to extract and display status updates.

Key mechanisms include:

  • Event Listening: It tracks tool_execution_start and tool_execution_update to show real-time tool names and progress percentages.
  • Model Narration: It injects a system prompt instruction requiring the model to start each step with a short description (e.g., ⏵ Checking error logs), which the extension then extracts and displays.
  • State Management: It manages a queue for fast-executing tools to ensure they are visible to the user and handles context window warnings by monitoring usage thresholds.

Who it’s for

Users of pi CLI and DeepSeek Harness who want a more transparent, engaging, and informative user interface while their AI agents are performing complex tasks.

Highlights

  • Real-time Tool Tracking: Displays actual tool activity and progress (percentages/stages) rather than fake animations.
  • Context Alerts: Warns users when context usage exceeds a 80% threshold (yellow) or 95% (red).
  • Model Self-Description: Forces the model to narrate its current action in under 20 words for real-time display.
  • Dynamic Content: Features a library of 95+ colloquial phrases, time-aware greetings (night/weekend), and rare "rainbow" easter eggs to make the agent feel more alive.
  • Performance Metrics: Provides post-task summaries including total time, tool count, cost (via official API usage), and token counts.
  • Git Integration: Automatically detects and displays the current git branch during git-related operations.
  • Customization: Includes a searchable settings panel, 28 animation presets, and a "Doctor" tool for configuration health checks.

Related

  • Project
    Aisland-SJL/dsh-worktable

    An agent-project workbench for DeepSeek Harness that provides a dockable multi-window workspace and a real-time control room for monitoring AI projects.

  • Project
    bowenliang123/dsh-context

    A DeepSeek Harness plugin that provides a comprehensive dashboard and tools for monitoring and managing an AI agent's context window, token usage, and file activity.

  • Project
    AMAP-ML/LongHorizon-Harness

    A loop engineering framework that turns existing AI agents into long-running computer-use systems by wrapping them in a structured plan-act-verify cycle across GUI and CLI environments.

  • Project
    joeynyc/hermes-hudui

    A browser-based monitoring dashboard for the Hermes AI agent that provides visibility into its memory, costs, and internal state.

  • Project
    NousResearch/hermes-agent

    A self-improving AI agent with a built-in learning loop that creates and refines skills from experience and maintains persistent user memory across sessions.