Agent skill

Mgsc Table Probe

by OpenXiangShan in OpenXiangShan/GEM5

运行或分析 MGSC 微测试的 SC 子表 A/B 实验,用分支与 MGSCTRACE 统计归因并改进测试. An agent skill from OpenXiangShan/GEM5.

BSD-3-ClauseAuto-check passed

Install Mgsc Table Probe

skills CLI
$ npx skills add OpenXiangShan/GEM5 --skill mgsc-table-probe -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install OpenXiangShan/GEM5 mgsc-table-probe --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/OpenXiangShan/GEM5.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/mgsc-table-probe .claude/skills/mgsc-table-probe && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
mgsc-table-probe
GitHub stars
162
Token cost
~628 tokens
SKILL.md length
160 words
Files
5 (incl. scripts, references)
Skills in repo
10
Repo updated
First seen
Licence
BSD-3-Clause

At a glance

运行或分析 MGSC 微测试的 SC 子表 A/B 实验,用分支与 MGSCTRACE 统计归因并改进测试. An agent skill from OpenXiangShan/GEM5.

  • Works in 3 steps: 对所有 mgsc_test 二进制执行 probe → 快速检查单个 case → 仅重建报告(不重新运行 gem5)
  • SKILL.md covers 概览, 快速开始, 路径解析 and 工作流, plus 4 more sections
  • Runs Python scripts from its folder; calls python3

What it does

Mgsc Table Probe is an agent skill from OpenXiangShan/GEM5. 运行或分析 MGSC 微测试的 SC 子表 A/B 实验,用分支与 MGSCTRACE 统计归因并改进测试。

Its SKILL.md is about 630 tokens, which your agent loads only when the skill is triggered. The skill folder holds 7 other files, including scripts and reference files (for example `agents/openai.yaml`, `references/test-patterns.md` and `scripts/mgsc_table_probe.py`).

The licence is BSD-3-Clause.

Example prompts

  • “/mgsc-table-probe”

Requirements

  • Python 3

Workflow steps

3 steps, taken from the first numbered list in SKILL.md.

  1. 对所有 mgsc_test 二进制执行 probe
  2. 快速检查单个 case
  3. 仅重建报告(不重新运行 gem5)

What it can do on your machine

Read from SKILL.md and the folder at commit 0785475. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Ships 2 files in scripts/ (Python), which the agent can run.

    Shell commands in SKILL.md call:

    • python3

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Mgsc Table Probe loads about 628 tokens when it runs, and up to ~1.2k if it reads all its reference files. Until then it costs about 18 tokens; SKILL.md has 160 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~18
When it runs · the whole SKILL.md, loaded when a task matches
~628
With references · SKILL.md plus every file in references/, read only if the agent opens them
~1.2k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from OpenXiangShan/GEM5 at commit 0785475, republished under its BSD-3-Clause licence (© OpenXiangShan). 160 words, ~628 tokens.

Download SKILL.mdSave it as .claude/skills/mgsc-table-probe/SKILL.md (or your agent's skills folder). This skill also uses 4 other files; get the full folder from GitHub.
name
mgsc-table-probe
description
运行或分析 MGSC 微测试的 SC 子表 A/B 实验,用分支与 MGSCTRACE 统计归因并改进测试。

MGSC 表探测

概览

运行标准化的 SC 表 A/B 实验,并产出分支级别的归因结果:

  • summary.csv:case 级别的 off 与各 profile 的 delta 汇总。
  • branch_delta.csv:分支级别的误预测 delta,以及 SC 修复/伤害与表贡献比例。
  • report.md:便于人工快速阅读的排序报告。

这个 skill 的目标是快速迭代 SC 测试质量,而不是做完整的性能调优。

快速开始

  1. 对所有 mgsc_test 二进制执行 probe:
bash
python3 .agents/skills/mgsc-table-probe/scripts/mgsc_table_probe.py \
  --outdir debug/sc_table_probe \
  --profiles off,l_only,g_only,i_only,full \
  --max-workers 4
  1. 快速检查单个 case:
bash
python3 .agents/skills/mgsc-table-probe/scripts/mgsc_table_probe.py \
  --outdir debug/sc_table_probe_smoke \
  --tests fp_sc_alias_pair \
  --profiles off,g_only,i_only \
  --max-workers 1
  1. 仅重建报告(不重新运行 gem5):
bash
python3 .agents/skills/mgsc-table-probe/scripts/mgsc_table_probe.py \
  --outdir debug/sc_table_probe \
  --profiles off,l_only,g_only,i_only,full \
  --skip-run

路径解析

  • GEM5 仓库:优先使用 GEM5_HOME,其次从当前目录和脚本位置向上查找 SConstruct 与 configs/example/kmhv3.py。
  • MGSC 测试:优先使用 AM_HOME;未设置时才采用本机默认 /nfs/home/yanyue/tools/nexus-am。
  • 所有路径都可用 --gem5-bin、--config、--cpt-dir 和 --src-dir 显式覆盖。

个人 home 路径只是方便本机使用的 fallback,不是跨机器前提。脚本会在创建运行任务前检查 checkpoint 目录、gem5 binary 和配置文件。

工作流

  1. 基线 + 单表隔离 profile
  • 始终包含 off。
  • 在检查 full 之前,先加入单表 profile(如 g_only、i_only)。
  1. 快速筛选有价值的测试
  • 在 summary.csv 中,优先关注满足 condMiss_delta < 0 且 mgsc_net_use > 0 的 case。
  1. 筛选有价值的分支
  • 在 branch_delta.csv 中,优先关注满足以下条件的行:
    • delta_misp < 0
    • focus_decisive_ratio 较高
    • focus_agree_fix_ratio 较高
  1. 决定下一步微测试方向
  • 如果 g_only 很少带来改善,且 focus_decisive_ratio(g) 很低,说明 global-history 模式较弱。
  • 如果 i_only 从未带来帮助,说明循环/迭代相位信号暴露得还不够。
  • 可以参考 references/test-patterns.md 里的模式来编写下一个测试。

输出

  • debug/sc_table_probe/summary.csv
  • debug/sc_table_probe/branch_delta.csv
  • debug/sc_table_probe/report.md
  • debug/sc_table_probe/report.json

关键参数

  • --profiles:从 off,l_only,g_only,i_only,full 中选择。
  • --tests:逗号分隔的测试名(不带后缀),例如 fp_sc_alias_pair,imli_iter。
  • --extra-param:透传额外的 gem5 --param。
  • --copy-cpt-to-tmp:避免路径访问问题。
  • --skip-run:仅生成报告,不执行运行。

注意事项

  • 除非你明确要评估交互效应,否则在做 SC 子表归因时应保持 microtage 关闭。
  • 不同 profile 之间必须使用同一组 checkpoint,否则 delta 无效。
  • 做 branch PC 映射时,使用 mgsc_test 构建目录下的 *-riscv64-xs.txt 反汇编文件。

参考资料

  • 关于面向 G/IMLI 的微测试模式,参见 references/test-patterns.md。
  • scripts/mgsc_trace_report.py 可用于对已有 bp.db 做更细的 branch/run 级门限与表贡献报告;它不负责重跑 gem5。

© OpenXiangShan, BSD-3-Clause. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 4 other files (scripts, references) in .agents/skills/mgsc-table-probe of OpenXiangShan/GEM5.

  • SKILL.md
  • agents/openai.yaml
  • references/test-patterns.md
  • scripts/mgsc_table_probe.py
  • scripts/mgsc_trace_report.py

Open the folder on GitHubat commit 0785475

Compare with similar skills

Mgsc Table Probe next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Mgsc Table Probe compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Mgsc Table Probe this skillOpenXiangShan/GEM5162—~628Automated safety check: PassBSD-3-Clause
Table Fitasgeirtj/system_prompts_leaks69k—~772Automated safety check: PassCC0-1.0
Table Headersthedaviddias/Front-End-Checklist74k—~460Automated safety check: PassMIT
Accessible Tablesthedaviddias/Front-End-Checklist74k—~521Automated safety check: PassMIT
Data Table Managern8n-io/n8n207k—~2.3kAutomated safety check: PassCustom licence
Tablebrycewang-stanford/Auto-Empirical-Research-Skills4.6k—~3.3kAutomated safety check: PassCustom licence

Similar skills

  • Table Fit

    asgeirtj/system_prompts_leaks

    Keep a Markdown table readable in a narrow terminal of about 100 display columns - a wide table or one carrying prose in its cells wraps into unreadable ragged rows.

    69k GitHub stars~772 tokensUpdated today
    Testing & QAAuto-check passed
  • Table Headers

    thedaviddias/Front-End-Checklist

    A skill your agent uses when reviewing rendered HTML, interactive components, or design-system patterns related to Define proper table headers.

    74k GitHub stars~460 tokensUpdated 4 days ago
    Frontend & DesignAuto-check passed
  • Accessible Tables

    thedaviddias/Front-End-Checklist

    A skill your agent uses when reviewing rendered HTML, interactive components, or design-system patterns related to Use semantic table markup for screen readers.

    74k GitHub stars~521 tokensUpdated 4 days ago
    Frontend & DesignAuto-check passed
  • Official

    Load before calling data-tables or parse-file. An agent skill from n8n-io/n8n.

    207k GitHub stars~2.3k tokensUpdated today
    Documents & OfficeAuto-check passed
  • Table

    brycewang-stanford/Auto-Empirical-Research-Skills

    Econometrics skill for creating publication-quality LaTeX regression and summary tables.

    4.6k GitHub stars~3.3k tokensUpdated 5 days ago
    Documents & OfficeAuto-check passed
  • Tables

    brycewang-stanford/Auto-Empirical-Research-Skills

    Design and format publication-quality tables: column order, row grouping, notes, precision, reproducibility.

    4.6k GitHub stars~2k tokensUpdated 5 days ago
    Research & ScienceAuto-check passed

More from OpenXiangShan/GEM5

All 10 skills in this repo
  • Tage Trace Workflow

    OpenXiangShan/GEM5

    采集或分析 gem5/XiangShan TAGE trace,比较统计、热点分支和事件分叉. An agent skill from OpenXiangShan/GEM5.

    162 GitHub stars~629 tokensUpdated today
    Auto-check passed
  • CI Perf Analysis

    OpenXiangShan/GEM5

    定位 GEM5 性能 CI 的归档与分数,比较 run 的 benchmark 表现并用 stats 分析变化. An agent skill from OpenXiangShan/GEM5.

    162 GitHub stars~545 tokensUpdated today
    Auto-check passed
  • Frontend Pmu Analysis

    OpenXiangShan/GEM5

    仅做 BPU 计数器提取与批量汇总(机器可读 JSON/CSV)。配置文件只需要写原始 stats 计数器名. An agent skill from OpenXiangShan/GEM5.

    162 GitHub stars~305 tokensUpdated today
    Auto-check passed
  • CI Param Solver

    OpenXiangShan/GEM5

    编写或校验 GEM5 SolveSpec 参数搜索,并按请求准备或触发 manual-solve.yml. An agent skill from OpenXiangShan/GEM5.

    162 GitHub stars~389 tokensUpdated today
    Auto-check passed
  • CI Perf PR Report

    OpenXiangShan/GEM5

    为 GEM5 PR 准备或发布性能对比报告,选择可比 baseline 并汇总分数与关键计数器证据. An agent skill from OpenXiangShan/GEM5.

    162 GitHub stars~462 tokensUpdated today
    Auto-check passed
  • Gem5 Manual Perf Trigger

    OpenXiangShan/GEM5

    用于在本地通过 gh 远程触发 OpenXiangShan/GEM5 的 manual-perf.yml,并在触发后检查 run 是否正常创建。适用于用户明确要求触发 manual-perf、组装 gh workflow run 命令、校验 workflow 输入项、或确认触发后的 Actions run 状态。

    162 GitHub stars~763 tokensUpdated today
    Auto-check passed

Questions about Mgsc Table Probe

What does Mgsc Table Probe do?

运行或分析 MGSC 微测试的 SC 子表 A/B 实验,用分支与 MGSCTRACE 统计归因并改进测试. An agent skill from OpenXiangShan/GEM5. Mgsc Table Probe is an agent skill from OpenXiangShan/GEM5.

How do I install Mgsc Table Probe in Claude Code?

Run `npx skills add OpenXiangShan/GEM5 --skill mgsc-table-probe -a claude-code`. Or copy the skill folder (.agents/skills/mgsc-table-probe in OpenXiangShan/GEM5) into .claude/skills/mgsc-table-probe in your project. Claude Code loads it when a task matches its description.

How do I install Mgsc Table Probe in Codex?

Run `npx skills add OpenXiangShan/GEM5 --skill mgsc-table-probe -a codex`. Or copy the skill folder (.agents/skills/mgsc-table-probe in OpenXiangShan/GEM5) into .agents/skills/mgsc-table-probe in your project. Codex loads it when a task matches its description.

Can I use Mgsc Table Probe in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add OpenXiangShan/GEM5 --skill mgsc-table-probe -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/mgsc-table-probe, .gemini/skills/mgsc-table-probe, .github/skills/mgsc-table-probe and .opencode/skills/mgsc-table-probe in your project.

What does Mgsc Table Probe need to run?

Going by SKILL.md and its folder, Mgsc Table Probe needs Python for the scripts in its folder and the command-line tools its instructions call (python3). Our summary lists: Python 3.

Does Mgsc Table Probe access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Mgsc Table Probe safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.

What licence does Mgsc Table Probe use?

Mgsc Table Probe is published under the BSD-3-Clause licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Mgsc Table Probe use?

About 628 tokens (SKILL.md is roughly 2.5k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 523 tokens, read only when the agent opens those files.

What are the alternatives to Mgsc Table Probe?

Skills that share tags, products or a category with Mgsc Table Probe: Table Fit (asgeirtj/system_prompts_leaks, 69k stars), Table Headers (thedaviddias/Front-End-Checklist, 74k stars), Accessible Tables (thedaviddias/Front-End-Checklist, 74k stars) and Data Table Manager (n8n-io/n8n, 207k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Mgsc Table Probe?

OpenXiangShan (a GitHub organization) maintains it in OpenXiangShan/GEM5, which has 162 GitHub stars. The repository holds 10 skills in this directory. The repository was last updated on October 10, 2026.

Source: OpenXiangShan/GEM5 on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.