registry / @tea-agent/loop-agent / 0.2.0

@tea-agent/loop-agent@0.2.0

已弃用：此版本可能包含过时内容，请升级到最新版 npm i @tea-agent/loop-agent@latest

`loop-agent` 是面向 AI coding agent 的仓库级任务运行时和治理工具。它把一次研发任务组织成可生成、可校验、可执行、可恢复、可交接的 Agent DAG，并用 `.harness/`、`docs/` 和 shell verification 记录执行事实、长期治理资料和完成依据。

AI Security Review

scanned 8d ago · by lpm-firewall-ai

No confirmed malicious attack surface. The CLI intentionally runs configured Pi/Cursor/shell workflows and can initialize files beneath an explicit target repository, but neither occurs at installation or import time.

Static reason

No blocking static signals were detected.

Trigger

User invokes loop-agent commands such as init, run-dag, pi-prompt, or cursor-prompt.

Impact

Configured commands may modify the selected project or invoke configured AI tooling; no stealth persistence, credential exfiltration, or unconsented foreign agent-control mutation is established.

Mechanism

User-invoked agent orchestration, child-process execution, and repo-local harness generation.

Rationale

Static source inspection shows a user-invoked coding-agent CLI with expected filesystem and subprocess capabilities. It has no install-time execution or concrete exfiltration, remote payload, stealth persistence, or foreign/broad AI-agent control-surface mutation chain.

Evidence

package.jsonbin/loop-agent.jsdist/commands/init.jsdist/executors/pi-executor.jsdist/executors/cursor-executor.jsdist/executors/shell-executor.jsdist/workflows/dag/canvas-observer.jsdist/cli.js

Decision evidence

public snapshot

AI called this at 91.0% confidence with false-positive risk.

@tea-agent/loop-agent@0.2.0

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code