惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

V
Visual Studio Blog
J
Java Code Geeks
H
Hackread – Cybersecurity News, Data Breaches, AI and More
D
Docker
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
博客园 - 聂微东
MyScale Blog
MyScale Blog
H
Help Net Security
Last Week in AI
Last Week in AI
T
The Blog of Author Tim Ferriss
M
MIT News - Artificial intelligence
大猫的无限游戏
大猫的无限游戏
酷 壳 – CoolShell
酷 壳 – CoolShell
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
P
Proofpoint News Feed
博客园 - 叶小钗
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
Y
Y Combinator Blog
Recent Announcements
Recent Announcements
F
Fortinet All Blogs
Martin Fowler
Martin Fowler
Microsoft Security Blog
Microsoft Security Blog
T
Tailwind CSS Blog
aimingoo的专栏
aimingoo的专栏

Hacker News - Newest: "AI"

AI can't read an investor deck AI as an attorney? Student uses ChatGPT, Gemini to sue UW over alleged racial discrimination Hacking MCP Servers in AI Systems – The Rug Pull: Tool Changes After Approval GitHub - MeepCastana/KubeezCut: Free Web based video editor Can AI judge journalism? A Thiel-backed startup says yes, even if it risks chilling whistleblowers Coming soon: 10 Things That Matter in AI Right Now DARPA built an AI to fact-check enemy weapons claims What explains heterogeneity in AI adoption? When AI Meets Muscle: Context-Aware Electrical Stimulation Promises a New Way to Guide Human Movements - Department of Computer Science AI Changed How We Build. It Did Not Change What Matters. Linux rules on using AI-generated code - Copilot is OK, but humans must take 'full responsibility for the… Meta spins up AI version of Mark Zuckerberg to engage with employees Code Mode: Let Your AI Write Programs, Not Just Call Tools | TanStack Blog GitHub - Delavalom/graft: Go framework for building AI agents. Type-safe tools, multi-provider (OpenAI, Anthropic, Gemini, Bedrock), zero vendor SDKs. India's TCS tops estimates, says new AI models did not dent services demand Gen Z's fading AI hype Strong feeling: we are in a folded AI reality GitHub - machinarii/total-recall-catalog: A reference catalog of latest knowledge retrieval, memory & RAG systems GitHub - mensfeld/code-on-incus: Give each AI agent its own isolated machine with root, Docker, and systemd. Active defense detects and stops threats automatically.. Quantization, LoRA, and the 8% Problem: Benchmarking Local LLMs for Production AI Iran war: We spoke to the man making Lego-style AI videos that experts say are powerful propaganda Powell, Bessent discussed Anthropic's Mythos AI cyber threat with major U.S. banks GitHub - immartian/bellamem: Persistent belief-graph memory for AI agents. Retrieves decisive context by importance — not recency, not RAG, not /compact. recursive-mode: The Repo-Native Operating System for AI Engineering After the attack on Sam Altman's home, will AI CEO's go on the offensive? The biggest advance in AI since the LLM Opus 4.6 vs GPT 5.4 One Prompt Unity World Generation Test “AI polls” are fake polls Client Challenge Can AI be a 'child of God'? Inside Anthropic's meeting with Christian leaders
GitHub - HoratiuCode/Clawie: Clawie is a fast terminal co...
horatiucode · 2026-06-27 · via Hacker News - Newest: "AI"

Clawie

Clawie by ShrimpAI

Clawie is a packaged workspace that keeps both sides of the project together:

  • rust-clawie (main CLI/runtime side)
  • python-clawie (Python mirror/workspace side)

This repository exists so the full project is shareable and runnable from one folder.

Key Features & Recent Updates

1. Interactive Setup Wizard

Configure your environment variables, provider, model, API key, and base URL interactively:

Settings are persisted in settings.json under the Clawie config directory.

2. Lazy Senior Dev Mode (Lean Mode)

Clawie enforces a Lean Ladder to prevent over-engineering:

  1. Does this need to exist?
  2. Is it already in this codebase?
  3. Does stdlib do it?
  4. Does a native platform feature cover it?
  5. Does an installed dependency solve it?
  6. Can it be one line?
  7. Only then write the minimum code.

Manage this mode directly inside the REPL session:

  • /lean [lite|full|ultra|off]: Switch or view the active lean mode (default is full).
  • /lean-review: Review current diff for over-engineering.
  • /lean-audit: Scan repository for over-engineering.
  • /lean-debt: Harvest clawie: simplification comments into a ledger.
  • /lean-gain: Show benchmark impact metrics.
  • /lean-help: Print command reference.

3. Repository Mapping

Use the /map (or /repo-map) command in the REPL to generate a ranked map of the repository's files and extracted symbols, helping navigate large codebases.

4. Git Integration

Manage commits directly from the REPL:

  • /commit: Preflight checks changes, generates a commit message, and commits them.
  • /undo: Undoes the last commit (soft reset, keeping changes).

5. Workspace RAG Service (claw-rag-service)

SQLite-backed vector indexing service for semantic repository searches:

  • Ingest files: cargo run -p claw-rag-service -- ingest --workspace .
  • Serve API & UI: cargo run -p claw-rag-service -- serve

6. Web UI Visual Upgrades

Advanced graphical interface features for the local Clawie Web UI:

  • WebSocket Live Log Streaming: Dynamic real-time execution log streams. Rather than pulling static snapshots, the UI connects to a background socket (/ws-log) to monitor process events as they happen.
  • Side-by-Side Visual Diffing: Compare original files vs agent improvements or current edits. Clicking "Show Diff" provides visual red/green deletions/additions side-by-side with automatic layout alignment.

Clawie Web UI Overview

7. Automated Parity Pipelines

Checks sync and parity between the Rust codebase and Python mirrors:

  • Sync Auditor CLI: ./scripts/check_rust_python_sync.py analyzes command/tool definitions and file parity.
  • Unit Testing: Tests defined in test_rust_python_sync.py run checks in continuous integration.

8. Pixel Agents Dashboard (Visual Interface)

A gamified, real-time pixel-art dashboard showing active agent instances and status:

  • Draggable Agents: CLI processes are rendered as active pixel-art characters in visual rooms (complete with desks, computers, bookshelves, and server racks).
  • Session Actions: Terminate active agent sessions directly from the visual interface.
  • State Beacons: Displays process statuses (thinking, executing, idle, closed) dynamically via color-coded status lights.

Pixel Agents Visualizer

Quick Start

  1. Set up the workspace:
  1. Launch the Clawie agent REPL:
  1. Work in these folders depending on focus:
  • rust-clawie for CLI/runtime behavior
  • python-clawie for Python-side mirrored modules and tooling

Repository Layout

Clawie/
├── README.md
├── clawie
├── rust-clawie/
└── python-clawie/

Long Coding Sessions (Improved)

The runtime defaults were increased to better support longer sessions:

  • max_turns: 64 (was lower)
  • max_budget_tokens: 12000 (was lower)
  • compact_after_turns: 48
  • turn-loop --max-turns: default 12

You can tune these at runtime with environment variables:

export CLAWIE_MAX_TURNS=120
export CLAWIE_MAX_BUDGET_TOKENS=30000
export CLAWIE_COMPACT_AFTER_TURNS=80
export CLAWIE_STRUCTURED_OUTPUT=false
export CLAWIE_STRUCTURED_RETRY_LIMIT=2
./clawie

Notes:

  • Invalid values fall back to defaults.
  • Numeric values are clamped to at least 1.

Useful Commands

Run from repository root.

# Python-side summary
python3 -m python-clawie.src.main summary

# Run a stateful loop with explicit turn count
python3 -m python-clawie.src.main turn-loop "audit this module" --max-turns 30

# Resume an existing session
python3 -m python-clawie.src.main resume-session <session_id> "continue"

Use Cases & Workflows

1. Live Background Agent Monitoring (WebSocket Logs)

When running complex agent tasks (e.g. running a multi-turn audit session with /lean-audit), developers can launch the Web UI alongside their terminal and watch the agent's actions live.

  • Workflow:
    1. Start a CLI session: ./clawie
    2. Open the Web UI by running /webui in the CLI REPL or by running ./clawie --webui
    3. Click on the active room's terminal monitor inside the Web UI dashboard to open the Log Console.
    4. The console connects via WebSocket to /ws-log?pid=<PID> and streams process lifecycle updates, command elapsed times, and execution details in real time as the agent runs.

2. Code Improvement Review (Side-by-Side Diff)

When Clawie suggests changes to code files, you can review, edit, and apply them using the side-by-side split screen.

  • Workflow:
    1. Ask Clawie to improve a file: "Optimize main.py" (which generates a .improvements.md file).
    2. Open the Web UI and select the file from the workspace explorer sidebar.
    3. Click the "Show Diff" button at the top right of the editor.
    4. Compare the Original File (left pane, red deletions) and the Improvements / Edited (right pane, green additions).
    5. Switch back to the editor with "Show Editor" to make manual refinements, then click "Save".

3. Continuous Parity Checking (Automated Pipelines)

To ensure that CLI runtime command/tool updates inside rust-clawie are mirrored properly inside python-clawie without creating feature drift:

  • Workflow:
    1. Run the sync check CLI tool: ./scripts/check_rust_python_sync.py
    2. The tool outputs a detailed parity report comparing commands in Rust commands.json vs Python's snapshot, and tool specifications in Rust tools/src/lib.rs vs Python's snapshot.
    3. If there are missing files or content drifts, the script exits with code 1, serving as a validator in Git hooks or CI pipelines.
    4. Run python3 -m unittest python-clawie/tests/test_rust_python_sync.py to assert package structure sync.

Product Naming

  • Clawie: product name
  • ShrimpAI: parent brand
  • Jameclaw: legacy/origin naming context

Why This Package Exists

Earlier working copies were split across multiple local folders. This package keeps everything in one Git-ready structure so onboarding, development, and sharing are simpler.

ShrimpAI mascot