Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add ellmos-ai/skills --skill zhgit clone --depth 1 https://github.com/ellmos-ai/skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/ellmos-ai/skills/zh)<a href="https://agentmods.dev/skills/ellmos-ai/skills/zh"><img src="https://agentmods.dev/badge/skills/ellmos-ai/skills/zh/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/ellmos-ai/skills/zh"><img src="https://agentmods.dev/badge/skills/ellmos-ai/skills/zh.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00044 | $0.02764 |
| Opus 5 | $0.00022 | $0.01382 |
| Sonnet 5 | $0.00009 | $0.00553 |
| Haiku 4.5 | $0.00004 | $0.00276 |
Grade C, and why
bugfix-protocol scanned grade C with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
find . -name "__pycache__" -type d -exec rm -rf {} + 2>&1 This is a copy
86% identical to bugfix-protocol — 298 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 334 lines — stays where its author put it; the contents beside it link to each section on GitHub.
中文 —
bugfix-protocol官方中文版本。
Bugfix Protocol: 系统的 6 阶段 Debugging 协议
结构化的 Bug 处理方法 — 从症状分析到验证。 防止盲目的试错，确保修复方案可持续。
概述与目的
| 阶段 | 名称 | 目标 | 最长时间 |
|---|---|---|---|
| 1 | 快速检查 | 排除明显原因 | 2 分钟 |
| 2 | 诊断 | 定位根本原因 | 10 分钟 |
| 3 | 隔离测试 | 使 Bug 可复现 | 5 分钟 |
| 4 | 修复 | 最小化修正 | 10 分钟 |
| 5 | 验证 | 验证修复 + 检查副作用 | 5 分钟 |
| 6 | 文档编写 | 保存知识经验 | 2 分钟 |
20分钟规则： 如果 20 分钟后仍无进展，请改变方法或寻求帮助。
阶段 1: 快速检查 (2 分钟)
在深入分析之前 — 检查最常见的原因：
检查清单
- 语法错误？ 仔细阅读错误信息，检查对应行
- 导入错误？ 模块已安装？名称正确？循环导入？
- 拼写错误？ 变量/函数名称是否正确？
- 数据类型错误？ 字符串误作为整型？本应为对象的地方为 None？
- 缓存过期？ 删除
__pycache__，重新启动 - 环境错误？ 是否激活了正确的 venv？Python 版本是否正确？
- 编码问题？ UTF-8 vs. cp1252 (Windows 经典)
快速操作
# 清除缓存
find . -name "__pycache__" -type d -exec rm -rf {} + 2>&1
find . -name "*.pyc" -delete 2>&1
# 检查导入
python -c "import modulename"
# 检查语法
python -m py_compile file.py
阶段 2: 诊断 (10 分钟)
策略：由外入内 (Outside-In)
- 分析错误信息 — 从下往上阅读堆栈追踪 (traceback)
- 检查最近的修改 —
git diff,git log --oneline -10 - 使用诊断工具 — 使用项目专属的诊断工具
诊断工具 (示例)
根据项目的不同，专用的诊断脚本可能会大有帮助：
| 工具 | 用途 |
|---|---|
import_diagnose.py |
分析导入问题 |
method_analyzer.py |
检查方法签名 |
env_checker.py |
验证环境变量/路径 |
注意： 创建项目专属的诊断工具或使用现有的工具。 重要的是系统性的方法，而不是具体的工具。
调试技巧
# 1. Print 调试 (简单但高效)
print(f"DEBUG: variable={variable!r}, type={type(variable)}")
# 2. 断点 (交互式)
breakpoint() # Python 3.7+
# 3. 详细堆栈追踪
import traceback
traceback.print_exc()
# 4. 使用日志替代 print
import logging
logging.basicConfig(level=logging.DEBUG)
logger = logging.getLogger(__name__)
logger.debug(f"State: {state!r}")
阶段 3: 隔离测试 (5 分钟)
最小可复现示例 (MRE)
目标：用最少的代码复现 Bug。
# test_bug.py — 最小复现测试
"""
Bug: [简短描述]
Expected: [预期发生的结果]
Actual: [实际发生的结果]
"""
# 最小化设置
# ... 仅保留核心必要代码
# Bug 触发器
# ... 触发 Bug 的精确代码
# 预期结果
# assert result == expected, f"Got {result}"
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 9d ago First seen · 334 lines · 44 tokens per session scan C fe0fb16eff7c
bugfix-protocol is a skill published in the GitHub repository ellmos-ai/skills (4 stars, last pushed today), licensed MIT. It adds 44 tokens to every session and 2,764 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). It is 86% identical to bugfix-protocol, differing in 298 lines, and is treated as a copy.
Other skills, from other repositories
accelerate
Run PyTorch training across GPUs with minimal changes.
jupyter-notebook
Iterative Python via live Jupyter kernel (hamelnb).
stack-trace-python-probe
Internal helper for meta-stack-trace-investigator. Use when a Python traceback needs Python-specific root-cause checks, pytest reproducer guidance, and defensive patch targets.
pyghidra-scripting
Write and run Python (PyGhidra) code inside the Ghidra session that ReVa's MCP server is already attached to, using the five ReVa scripting tools — run-script, list-scripts, read-script, write-script, edit-script. Use this whenever the user asks to execute Python against the current program, reach for the Ghidra Flat…
ctf-kit
Use when the user is solving an authorized CTF / lab reverse-engineering challenge focused on Windows application authentication or license-check bypass. Guides triage → static analysis → dynamic experiment planning → bypass verification, with strong evidence discipline and VM safety boundaries. NOT for real-world…
memory-lint
Use when the user wants to lint a Claude Code memory directory (/.claude/memory or custom path) for index inconsistency, broken cross-links, stale project state, duplicate / conflicting feedback rules, naming violations, frontmatter gaps, and oversized files. Phase 1 is a read-only scan and report. Phase 2 applies…