← 返回 Skills 市场
Phoenix Iterate
作者
tltby12341
· GitHub ↗
· v1.0.1
· MIT-0
166
总下载
0
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install phoenix-iterate
功能描述
AI-driven quantitative strategy iteration workflow — a complete loop of briefing, hypothesis, code generation, constitutional scan, backtest submission, fore...
安全使用建议
This skill appears internally consistent for orchestrating quantitative strategy iterations. Before installing: (1) review the full orchestrator.py (the provided snippet is large but truncated here) to confirm it makes no unexpected network calls or subprocess invocations; (2) install and inspect the companion skills (backtest-poller and qc-deep-feature-forensics) because those will likely require credentials and perform networked backtests — do not hand over secrets unless you trust those companion packages; (3) be aware the skill will create and modify local files (memory/lessons.json, bans.json, results) under your project directory; (4) ensure pandas is available in your environment. If you need higher assurance, share the rest of orchestrator.py (or run a code review) so I can re-evaluate with the complete source.
功能分析
Type: OpenClaw Skill
Name: phoenix-iterate
Version: 1.0.1
The phoenix-iterate skill is a legitimate workflow orchestrator designed for AI-assisted quantitative strategy development. The core logic in orchestrator.py manages a local 'constitutional memory' to track trading lessons and enforce code restrictions, while SKILL.md provides a disciplined iteration loop for the agent. There are no signs of malicious intent, data exfiltration, or dangerous execution patterns; the tool strictly handles local strategy files, backtest results, and configuration data as described.
能力评估
Purpose & Capability
Name/description match the actual behavior: the skill coordinates strategy briefings, code scans, backtest submission/monitoring (via a companion skill), diagnostics, and lesson recording. Required binary (python3) is appropriate. It declares the companion skills it depends on, which is coherent for backtest submission and forensic analysis.
Instruction Scope
SKILL.md instructs the agent to run the included orchestrator.py subcommands, run a code scanner against strategy files, and invoke sibling CLIs (../backtest-poller/cli.py and ../qc-deep-feature-forensics/deep_forensics.py). Those actions are within the stated purpose. Note: the skill reads/writes local project files (memory_dir results, lessons.json, bans.json) and expects companion skills to perform networked backtests — review those companion skills' behavior and what data is sent externally.
Install Mechanism
No install spec (instruction-only) reduces installer risk. There is a requirements.txt (pandas>=1.5.0) but no automatic installer; the user must provide dependencies. No downloads or external installers are embedded in the skill.
Credentials
The skill requests no environment variables or credentials. It writes to local directories (./memory, ./results by default) which is appropriate for storing lessons and scan bans. It delegates any credentialed network access (backtest submission) to companion skills, so it itself does not require unrelated secrets.
Persistence & Privilege
always is false and the skill does not request elevated or persistent platform-wide privileges. It persists its own memory files (lessons.json, bans.json) within its configured memory_dir, which is consistent with its purpose.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install phoenix-iterate - 安装完成后，直接呼叫该 Skill 的名称或使用
/phoenix-iterate触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.1
**Added dependency documentation and integration instructions.**
- Documents the required companion skills (`backtest-poller`, `qc-deep-feature-forensics`) and their commands.
- Updates command references to use the appropriate skill locations and CLI wrappers.
- Adds explicit installation steps for dependencies.
- Clarifies which commands belong to which skill and how to adjust for different install locations.
- No functional changes to code; documentation change only.
v1.0.0
Phoenix Iterate 1.0.0 introduces a disciplined, AI-driven workflow for iterative quant strategy development:
- Provides a complete loop: briefing, hypothesis, code mutation, constitutional scan, staged backtesting, forensic diagnosis, and lesson recording.
- Enforces single-dimension iteration for clear attribution of results.
- Progressive validation stages (smoke, stress, medium, full) minimize wasted compute and catch issues early.
- Includes strict rule set for process integrity and accountability.
- Offers detailed command-line usage instructions for each workflow step.
元数据
常见问题
Phoenix Iterate 是什么？
AI-driven quantitative strategy iteration workflow — a complete loop of briefing, hypothesis, code generation, constitutional scan, backtest submission, fore... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 166 次。
如何安装 Phoenix Iterate？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install phoenix-iterate」即可一键安装，无需额外配置。
Phoenix Iterate 是免费的吗？
是的，Phoenix Iterate 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Phoenix Iterate 支持哪些平台？
Phoenix Iterate 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Phoenix Iterate？
由 tltby12341（@tltby12341）开发并维护，当前版本 v1.0.1。
推荐 Skills