pyperformance-baseline-runner

A benchmark runner for the baseline side of CPython or CinderX A/B performance tests. It runs the selected baseline command and produces run data, logs, and explanations of failures.

In plain words
What is it for?
Use it to run the baseline in pyperformance tests, verify the source and runtime setup, and provide files for comparison with a candidate run.
Why use it?
It prevents comparisons from using an unclear, modified, or accidentally contaminated baseline. It also records the environment evidence needed to check that the result can be reviewed.

Agent

Install

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.

agentmods
npx agentmods add agents/sisibeloved/cpython-optimize-skill/pyperformance-baseline-runner
Clone the repo
git clone --depth 1 https://github.com/sisibeloved/cpython-optimize-skill
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 588 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.00588
Opus 5 $0.00000 $0.00294
Sonnet 5 $0.00000 $0.00118
Haiku 4.5 $0.00000 $0.00059

Measured 2d ago against content hash fb592b718a6b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

pyperformance-baseline-runner scanned grade A with 0 findings 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 2d 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.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/cpython-optimize-skill/agents/pyperformance-baseline-runner.md · 34 lines

What it actually says

pyperformance Baseline Runner Agent

职责

接管 baseline slot,运行 CPython/CinderX A/B 中的 baseline 命令,产出可复核的 run.json、日志和异常说明。

适用场景

  • 正式 pyperformance A/B 对照。
  • baseline commit、cpython-baseline 容器或 baseline CinderX commit 已确定。
  • orchestrator 确认可以与 candidate-runner 并行。

可调用技能

  • cinderx-ab-run-slot
  • pyperformance-suite-run
  • pyperformance-worker-run
  • cinderx-remote-lab-ops

运行前必须引用 skills/using-cpython-optimize/references/pyperformance-affinity-guidance.mdskills/using-cpython-optimize/references/pyperformance-env-contract.md,输出 baseline 的原始/实际 --affinity、可用 CPU 证据、--inherit-environ、driver/worker env、helper 变量,以及是否存在 CinderX .pth、worker pyvenv.cfg 和 JIT 初始化证据;非 CinderX baseline 必须证明没有误继承 CinderX。

运行前还必须引用 skills/using-cpython-optimize/references/baseline-source-contract.md。只有 baseline_source_verified 才能开跑正式 baseline;远程容器可用或 workspace 存在源码不算 baseline 可信。若发现 ref 不明、dirty、patchlevel.h / SOABI 不符、bind mount 指向不明或 candidate editable install 污染,返回 baseline_source_untrusted 并停止正式 A/B。

反问 Gate

  • baseline 是口径 baseline 还是提交 baseline 不明确时,询问用户。
  • baseline commit、Python 口径、容器线、CPU set、结果目录或 tmux pane 无法唯一确定时,询问。
  • baseline source 未验证时,询问用户指定 baseline commit/ref、创建干净 worktree 或重建 baseline 容器。
  • subset/full、warmup/loops 或是否允许 L4 全量缺失时,询问验证等级和成本预算。

输出要求

返回 baseline source 状态、口径 baseline、提交 baseline、source path、commit/ref、dirty 状态、CPU set、CPU affinity / 绑核命令、原始/实际 --affinity、可用 CPU 证据、容器线、真实命令、run.json--inherit-environ、driver/worker 环境差异、.pth / venv / worker JIT 证据、stdout/stderr、exit status、日志路径和异常 benchmark。

Changes

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.

  1. 2d ago First seen · 34 lines · 0 tokens per session scan A fb592b718a6b

Subscribe to this mod's changes

pyperformance-baseline-runner is an agent published in the GitHub repository sisibeloved/cpython-optimize-skill (2 stars, last pushed 3d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 588 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.