1 · Zero total minutes
3,870
年度工作流总分钟数
Total workflow minutes for last year
2026-02-12 → 2026-08-07 · GitHub Actions Last-year snapshot
2026-02-12 → 2026-08-07 · GitHub Actions Last-year snapshot
NEXUS LIFE-CYCLE WORKFLOW VISUALIZER
这是 zero-entropy-lab 与 welcome-to-github 的双工作流可视化主页, 用一页展示年度用量, 运行效率, 失败成本, 步骤差异, 以及实际运行日志证据
A single-page visualizer for the zero-entropy-lab and welcome-to-github workflows, showing yearly usage, execution efficiency, failure cost, stage differences, and concrete run-log evidence
Annual Actions Load
年度负载对比Yearly load comparison
这里使用 GitHub Actions last year 数据, 1 是 zero, 2 是 welcome, 窗口 2026-02-12 → 2026-08-07
This section uses the GitHub Actions last-year metrics, 1 is zero, 2 is welcome, window 2026-02-12 → 2026-08-07
1 · Zero job runs
3,172
年度 job run 总数
Total job runs for last year
2 · Welcome total minutes
7,045
年度工作流总分钟数
Total workflow minutes for last year
2 · Welcome job runs
5,786
年度 job run 总数
Total job runs for last year
Compute Budget
计算预算Compute budget
Welcome 约消耗 Zero 的 1.82 倍 Actions 分钟, 两仓共享同一 NEXUS 生命周期骨架, 承担不同的责任边界
Welcome uses about 1.82x the Actions minutes of Zero — both repos share the same NEXUS life-cycle skeleton while carrying different responsibility boundaries
Efficiency And Failure Cost
效率与失败成本Efficiency and failure cost
两边平均排队时间都是 3s, 失败率都是 2%, 差异主要在年度负载和失败分钟数, zero 约 70 failed minutes, welcome 是 26 failed minutes
Both lanes have a 3s average queue time and 2% job failure rate, while the main difference is annual load and failed-job minutes, zero is about 70 failed minutes and welcome is 26
37sZero avg run
42sWelcome avg run
3s / 3sAvg queue
2%Zero fail rate
2%Welcome fail rate
70 / 26Failed minutes
Two Concrete Workflow Lanes
两个具体工作流Two concrete workflow lanes
这不是泛生态展示, 而是把 zero 和 welcome 两条 NEXUS CORTEX Life Cycle 直接展开成可读步骤
This is not a generic ecosystem page, it expands the zero and welcome NEXUS CORTEX Life Cycle workflows into readable stages
1
zero-entropy-lab
Integrity and State Core
Integrity and State Core
3,172job runs
37savg run
2%failure
Verify foundation
编译 30 后端文件 + 7 模块全部通过
Compile 30 backend files + 7 modules all OK
Document hygiene
维护 JSONL 知识库 · 1,687 记录
Maintain JSONL knowledge base · 1,687 records
Harvest signals
收割外部仓库情报 · 4 更新
Harvest external signals · 4 updated
Project mapping
投影文档→仓库关系 · 211 文档
Project doc→repo relations · 211 docs
Canonicalize + prune
规范化 888 实体 · 哈希链验证 · 保护路径裁剪
Canonicalize 888 entities · hash chain · prune protected
Rebuild + ingest
1,687 记忆重建 + 35 文件映射入 Cortex
1,687 memories + 35 files into Cortex
Ponder
382 节点 · 211 边 · 密度 1.15 · FRAGMENTED
382 nodes · 211 edges · density 1.15 · FRAGMENTED
Evolve
净化 + 每日演化 · 认知报告 + 任务简报
Purification + daily cycle · cognitive report
Re-verify + commit
全部 7 模块重验 → 11 文件 · +1,477 −268
7 modules re-verified → 11 files · +1,477 −268
2
welcome-to-github
Harvester and Public Surface
Harvester and Public Surface
5,786job runs
42savg run
2%failure
Verify foundation
编译 15 生命周期文件 + 34 测试全部通过
Compile 15 lifecycle files + 34 tests OK
Document hygiene
维护 JSONL 知识库 · 1,350 记录
Maintain JSONL knowledge base · 1,350 records
Harvest external facts
扫描外部 AI 仓库信号 · 全部已最新
Scan external AI repos · all current
Project mapping
投影文档→仓库关系 · 20 文档
Project doc→repo relations · 20 docs
Canonicalize + prune
规范化 692 实体 · 保护路径裁剪
Canonicalize 692 entities · prune protected
Rebuild + ingest
Schema 重建 + 16 文件映射入 Cortex
Rebuild from schema + 16 files into Cortex
Ponder
PageRank 中心性 · 结构直觉 · 自驱目标
PageRank centrality · structural intuition · self-driven goal
Evolve
AST 变异沙盒通过 · 信任网关活跃 · 量化仪表盘
AST sandbox passed · Trust Gateway · quantitative dashboard
Re-verify + commit
34 测试重验通过 → 4 文件 · +90 −15
34 tests re-verified → 4 files · +90 −15
Step Comparison Matrix
步骤对比矩阵Step comparison matrix
两边共同拥有相同的 NEXUS 生命周期阶段, 但实施方式不同 — 比较同一阶段在不同仓库如何执行
Both lanes share the same NEXUS life-cycle stages but implement them differently — comparing how each stage executes across repositories
Stage
1 · zero
2 · welcome
Compile & Verify
30 backend files compiled · 7 test modules all OK
15 lifecycle files compiled · 34 tests OK
Document hygiene
1,687 records · 11 files · 0 duplicates
1,350 records · 10 files · 0 duplicates
Harvest
0 failed · 4 updated
0 failed · 0 updated (all current)
Project
211 docs · 39 repos · 211 relations
20 docs · 15 repos · 20 relations
Canonicalize
888 entities · 799 relations · hash chain verified
692 entities · 658 relations · 0 dangling
Prune protected
aegis-cortex, ballast, index.html, assets/**
horizon-cortex, parallax
Rebuild
1,687 memories restored from knowledge fragments
from schemas + JSONL fragments
Ingest
35 files mapped into Cortex
16 files mapped into Cortex
Ponder
382 nodes · 211 edges · density 1.15 · FRAGMENTED
PageRank centrality · isolation risk · self-driven goal
Evolve
purification · daily cycle · cognitive report
AST sandbox passed · Trust Gateway · quantitative dashboard
Re-verify
7 modules all OK · foundation verification passed
34 tests re-verified · canonicalize + rebuild OK
Commit
11 files · +1,477 −268 · cognitive update
4 files · +90 −15 · archive + ledger
Run Log Evidence
运行日志证据Run-log evidence
以下数据来自 2026-08-07 最新一次两仓同步运行, 不是样例, 是可验证的公开记录
The following data is from the August 7, 2026 synchronized run — not a sample, but a verifiable public record
CURRENT · 2026-08-07
[zero] verify 30 backend files compiled, 7 test modules, all OK
[zero] hygiene 1,687 records, 11 knowledge files, 0 duplicates
[zero] harvest 0 failed, 4 updated
[zero] project 211 docs, 39 repositories, 211 relations mapped
[zero] canonicalize 888 entities, 799 relations, hash chain verified (0 broken)
[zero] prune protected: aegis-cortex · ballast · index.html · assets · scripts · styles
[zero] rebuild 1,687 memories restored from knowledge fragments
[zero] ingest 35 files mapped into Cortex
[zero] ponder 382 nodes, 211 edges, FRAGMENTED_3_ORPHANS, density 1.15
[zero] evolve purification complete, cognitive report + 2 missions generated
[zero] re-verify 7 test modules all OK, foundation verification passed
[zero] commit 11 files, +1,477 −268, cognitive update pushed
[welcome] verify 15 lifecycle files compiled, 34 tests OK
[welcome] hygiene 1,350 records, 10 knowledge files, 0 duplicates
[welcome] harvest 0 failed, 0 updated (all sources current)
[welcome] project 20 docs, 15 repositories, 20 relations mapped
[welcome] canonicalize 692 entities, 658 relations, 0 duplicates, 0 dangling
[welcome] prune protected: horizon-cortex · parallax
[welcome] rebuild from schemas + JSONL fragments
[welcome] ingest 16 files mapped into Cortex
[welcome] ponder PageRank centrality: NEXUS System, isolation risk, self-driven goal
[welcome] evolve AST sandbox passed, Trust Gateway active, quantitative dashboard rendered
[welcome] re-verify 34 tests re-verified, canonicalize + rebuild OK
[welcome] commit 4 files, +90 −15, archive + ledger updated
HISTORICAL ARCHIVE · 2026-06-28 (click to expand)
[zero] rebuild Reconstruction Complete, 49397 memories restored
[zero] ingest 42 files mapped into Cortex
[zero] ponder nodes 70, topology HIGHLY_STRUCTURED_ZERO_ORPHANS, graph density 3.54
[zero] tests tests OK, MCP and server checks passed
[zero] commit 10 files changed, 545 insertions, cognitive report created
[welcome] harvest transformers, nexent, astron-agent, vllm, dify, mediapipe, markitdown, python-genai
[welcome] ingest 14 files mapped into Cortex
[welcome] ponder isolation risk, structural intuition, PageRank centrality, self-driven goal
[welcome] evolve AST sandbox passed, Trust Gateway Active, quantitative ledger rendered
[welcome] commit 15 files changed, 230 insertions, archive and ledger updated
Zero-Basics Explanation
0基础解析Zero-basics explanation
这部分给第一次看到 GitHub Actions 的人看, 不预设 CI/CD 背景
This part is for people seeing GitHub Actions for the first time, no CI/CD background required
什么是 workflow
What is a workflow
workflow 是仓库里的自动流程, 通常放在 .github/workflows 目录, 它告诉 GitHub 什么时候运行, 运行哪些 job, 每个 job 做什么
A workflow is an automated process stored in a repository, usually under .github/workflows, it tells GitHub when to run, which jobs to run, and what each job should do
什么是 job run
What is a job run
job run 是一次具体执行, 比如 checkout, setup python, rebuild, ingest, test, commit 都可以作为步骤出现在一次 job run 里
A job run is one concrete execution, steps such as checkout, setup python, rebuild, ingest, test, and commit can appear inside a job run
什么是 total minutes
What are total minutes
total minutes 表示工作流在 runner 上消耗的总运行时间, 数字越高, 说明这一条 lane 年度执行负载越重
Total minutes means the total runner time consumed by workflows — the higher the number, the greater the yearly execution volume
为什么看 queue time
Why queue time matters
queue time 是 job 等待 runner 的时间, 这里两边都是 3s, 表示调度等待不是主要差异, 差异主要来自实际执行量
Queue time is how long a job waits for a runner, both lanes are 3s here, so scheduling delay is not the main difference, execution volume is
为什么看 failure rate
Why failure rate matters
failure rate 表示失败 job 的比例, 两边都是 2%, 说明稳定性同级, failed job usage 用来观察失败消耗的分钟成本
Failure rate is the share of failed jobs, both lanes are 2%, so reliability is similar, failed job usage shows the minute cost of failed work
怎么读这页
How to read this page
先看 1 和 2 的年度数字, 再看两条 lane 的步骤, 最后用日志控制台验证这些数字背后的真实行为
Start with the yearly numbers for 1 and 2, then read the two lanes, then use the log console to connect the metrics to concrete behavior