89 / 90

第 89 天:纠正、部分失败反馈与 Agent 辅助工程检查

把第 88 天的未达项转化为已验证的纠正,并在 Capstone Agent 项目上完成一个有边界的委派任务——辅助证据严格分离。

Agent 与智能体工作流 480 分钟有效研习(8 小时)
01 120 分钟

部分失败分类法与纠正协议

Agent 辅助工程检查

前置要求

第 88 天评分表与未达标台账。

学习目标

把每个低于 2 分的量规项分类为 concept_gap、slip、ambiguity 或 evidence_gap,并应用五步纠正协议:复述、定位、最小复现、修复、复测。

操作指引

1. 四种未达类型及其特征:
- concept_gap:完全无法产出方法(如列不出链式法则)。走映射的恢复日——重新学习后挑战*新的*变体任务。
- slip:方法正确但算术/细节出错(如符号错、差一)。直接复测同任务,无需完整恢复日。
- ambiguity:任务措辞不清。记录不清之处,复测前先澄清。
- evidence_gap:答案正确但未运行验证命令。补出命令输出即可,无需重学。
2. 对每行台账应用纠正协议:(a) 用自己的话复述任务;(b) 精确定位失败点;(c) 构建展示该失败的最小复现;(d) 只修这一点;(e) 复测并记录验证输出。
3. 反馈质量规则:每条自我批注必须引用失败的量规准则,而非只写“错了”。例:“PY-1 未达准则 tie-order:most_common 返回插入序;修复=按(-count, word)排序”。
4. 逐行判断:走完整恢复日(concept_gap)还是 20 分钟定向复习(slip、evidence_gap)。

自主演练与记录

产出 corrections_plan.json:为第 88 天每行台账标注 miss_type、未达量规准则、已做纠正步骤、复测命令或变体任务。

独立掌握标准

每个失败项都诚实分类并带 miss_type、未达准则、纠正步骤与复测计划。

权威源资源与权利证明

linking Journal publisher copyright (all rights reserved) - lawful external linking only

Improving Students' Learning With Effective Learning Techniques: Promising Directions From Cognitive and Educational Psychology

作者 / 机构: John Dunlosky, Katherine A. Rawson, Elizabeth J. Marsh, Mitchell J. Nathan, Daniel T. Willingham | 原始出处链接 ↗

对应章节 / 时间戳: Psychological Science in the Public Interest 14(1): 4-58; practice testing and distributed practice sections

验证版本与日期: Psychological Science in the Public Interest, 2013 (Verified: 2026-10-30)

入选理由:

Detailed comparative review rating practice testing and distributed practice as high-utility techniques while rating highlighting and rereading low-utility; anchors the retrieval, correction, and spaced-review design of Days 88-90.

来源前置条件:

None beyond general academic reading.

安全回退方案(离线弹性):

Curriculum restates the utility ratings and application conditions in original prose; every task remains runnable offline.

对比候选资源与选型分析 (2)
Pashler et al. Learning Styles: Concepts and Evidence ↗ compared_alternative [Publisher copyright (originally Wiley; now SAGE Publications)]

Important negative result on learning-styles matching but narrower scope; Dunlosky et al. rates a broader set of techniques directly usable for this curriculum.

Brown, Roediger & McDaniel: Make It Stick: The Science of Successful Learning ↗ compared_alternative [Publisher copyright (Harvard University Press)]

Strong trade-book synthesis of the same evidence base; the peer-reviewed review is preferred for citation precision and technique-by-technique utility ratings.

linking PSF License Version 2

Python Standard Library: Built-in Exceptions

作者 / 机构: Python Software Foundation | 原始出处链接 ↗

对应章节 / 时间戳: Concrete exceptions and Exception hierarchy

验证版本与日期: Python 3.12 (Verified: 2026-10-30)

入选理由:

Normative exception hierarchy used to justify the narrowest correct except clause and FileNotFoundError attributes such as filename and errno.

来源前置条件:

Basic Python syntax from Study Days 01-09

安全回退方案(离线弹性):

The drill reproduces FileNotFoundError and ValueError locally, and help(ValueError) documents the hierarchy offline.

对比候选资源与选型分析 (2)
W3Schools Python Exceptions ↗ compared_rejected_for_days10_12 [Proprietary]

Beginner-friendly but non-normative; omits the OSError attributes and hierarchy precision the diagnosis drills require.

Broader class-oriented chapter; the dedicated exceptions reference is the tighter fit for reading and diagnosing concrete error types during debugging checks.

02 120 分钟

定向纠正与复测(无辅助)

独立掌握检查(无 AI 辅助)

前置要求

第 89 天活动 1 的纠正计划;辅助条件:独立。

学习目标

在相同无辅助条件下重做第 88 天全部失败任务,用可运行命令验证纠正,并诚实复评——前后对比,而非覆盖。

操作指引

1. 每个 concept_gap 行:先完成映射的恢复活动,然后挑战*变体*任务——不是原题,让复测度量迁移而非背答案。
2. 每个 slip 行:重答原任务并把两次作答做 diff。
3. 必须运行并记录的验证命令:梯度答案→float64 下 h=1e-5 的双侧有限差分 (f(x+h)−f(x−h))/(2h);git 序列→在一次性克隆上执行并展示 git log --oneline;Python 函数→python -m unittest -v;参数计数→重算并与 projects/pretrain/artifacts/run_artifact.json 做 diff。
4. 记录 corrections.jsonl——每个复测项一行:{task_id, before, after, verify_command, verified: true|false}。
5. 仍低于 2 分的保持低分——记录具名恢复路径,绝不静默升级。

自主演练与记录

输出覆盖第 88 天全部台账行的 corrections.jsonl,每行含前后得分与实际捕获的验证输出。

独立掌握标准

每个复测项要么是命令验证过的纠正,要么是带具名恢复路径的明确未达——不得静默升级。

权威源资源与权利证明

linking MIT

CS231n: Backpropagation and Intuitions

作者 / 机构: Andrej Karpathy / Stanford University | 原始出处链接 ↗

对应章节 / 时间戳: Patterns in backward flow

验证版本与日期: Course notes (Verified: 2026-10-30)

入选理由:

Modular local-gradient interpretation (gate operations: add, multiply, max) and gradient flow intuition.

来源前置条件:

Multivariable calculus and basic Python.

安全回退方案(离线弹性):

Local gate definitions and circuit flow diagrams in the lesson provide full coverage.

对比候选资源与选型分析 (2)

Raw circuit view; CS231n notes are more structured with explicit gate derivations.

Fast.ai Practical Deep Learning for Coders ↗ compared_alternative [GPL-3.0]

Top-down approach; CS231n bottom-up modular view matches our first-principles curriculum.

linking CC BY-NC-SA 3.0 noncommercial share-alike

Pro Git, Second Edition

作者 / 机构: Scott Chacon and Ben Straub | 原始出处链接 ↗

对应章节 / 时间戳: Chapter 3 Git Branching and Section 7.7 Reset Demystified

验证版本与日期: Second edition (Apress, 2014) (Verified: 2026-10-30)

入选理由:

Freely licensed reference for branching, merging, conflict resolution, and safe recovery vocabulary used on Day 14.

来源前置条件:

Basic git status/log/diff vocabulary from earlier study days

安全回退方案(离线弹性):

git help <command> and the Git manual pages ship with Git and cover the same commands offline; every required workflow is stated locally in the lesson.

对比候选资源与选型分析 (2)
GitHub Skills interactive courses ↗ compared_rejected_for_day14 [MIT (varies by course)]

Interactive but requires a GitHub account and an external service; disposable local repositories keep the exercise self-contained.

Atlassian Git Tutorials ↗ compared_rejected_for_day14 [All Rights Reserved]

Vendor-owned prose with no permissive license; Pro Git covers the same ground under CC BY-NC-SA.

03 120 分钟

Agent 辅助工程检查:在 Capstone Agent 上的有边界委派任务

Agent 辅助工程检查

前置要求

第 89 天活动 2;辅助条件:agent_assisted。本项是 Agent 辅助工程检查——其证据绝不转化为独立掌握。

学习目标

在 Capstone 项目上证明合格的 Agent 指挥能力:编写含验收标准与权限范围的有边界规格,在一次性分支上应用变更,审查 diff 是否越界,运行确定性评测,并记录诚实结果。

操作指引

1. 在 projects/agent_learning 上选一个有边界任务,如给 scorer CLI 增加 --json 输出标志,或给 summarize-notes fixture 增加一条笔记。
2. 委派前先写规格:目标、Agent 可触碰的文件(如 cli.py 与一个测试文件)、验收标准(python -m unittest tests.projects.test_agent_learning 通过;新标志输出合法 JSON)、停止条件。
3. 把 Agent 的补丁应用到一次性分支(自己的克隆内 git checkout -b delegated-task 或临时副本);不得用于受保护工作树。
4. diff 审查清单:无声明范围外文件;未新增网络调用;无密钥凭据;测试真正覆盖新行为;错误路径保留。
5. 运行验证:项目自身测试/评测命令+你的验收检查。记录:规格文本、变更文件、diff 审查要点、验证命令与输出、Agent 迭代次数、你的判定(accepted / accepted-with-fixes / rejected)。

自主演练与记录

输出 assisted_check.json:{spec, files_changed[], diff_review_notes, verify_command, verify_output_summary, agent_iterations, verdict}。

独立掌握标准

端到端完成一个委派任务,含已审查 diff、已验证测试运行、诚实迭代计数与有据判定——标注为辅助证据。

权威源资源与权利证明

linking Open Access / CC BY 4.0

ReAct: Synergizing Reasoning and Acting in Language Models

作者 / 机构: Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik Narasimhan, Yuan Cao | 原始出处链接 ↗

对应章节 / 时间戳: Section 3 & 4 (ReAct Formulation and Synergy)

验证版本与日期: ICLR 2023 / arXiv:2210.03629v2 (Verified: 2026-10-30)

入选理由:

Foundational academic paper introducing the interleaved Thought-Action-Observation control loop in LLM agents, establishing formal interleaving of task reasoning and external tool execution.

来源前置条件:

Prompt engineering, token generation basics, and conversational interaction loops.

安全回退方案(离线弹性):

Complete algorithmic step traces, state loop diagrams, and control loop pseudocode are self-contained in curriculum materials.

对比候选资源与选型分析 (2)

Pure reasoning without environmental actions or external tool grounding; ReAct extends reasoning into an interactive environment action-observation loop.

Early neuro-symbolic modular router concept; ReAct offers simpler, prompting-native action-observation loop dynamics suitable for standard API models.

linking Open Access / CC BY 4.0

SWE-bench: Can Language Models Resolve Real-World GitHub Issues?

作者 / 机构: Carlos E. Jimenez, John Yang, Alexander Wettig, Shunyu Yao, Kexin Pei, Ofir Press, Karthik R. Narasimhan | 原始出处链接 ↗

对应章节 / 时间戳: Section 2.1 Benchmark Construction and Section 2.2 Task Formulation (FAIL_TO_PASS tests)

验证版本与日期: ICLR 2024 / arXiv:2310.06770v3 (Verified: 2026-10-30)

入选理由:

Defines the standard pattern this block teaches: a bounded issue statement paired with executable fail-to-pass tests as verifiable acceptance criteria for agent-produced changes.

来源前置条件:

Unit testing, issue descriptions, and repository evaluation vocabulary.

安全回退方案(离线弹性):

The acceptance-criteria pattern is exercised locally through the Ticket 22 fixture scorer and the Ticket 24 regression exercise without external datasets.

对比候选资源与选型分析 (2)
HumanEval: Evaluating Large Language Models Trained on Code ↗ compared_alternative [arXiv non-exclusive license / Open Access]

Function-level synthesis benchmark; SWE-bench models the repository-scoped issue-to-diff workflow that this delivery block rehearses.

Multi-SWE-bench: A Multilingual Benchmark for Issue Resolving ↗ compared_alternative [Open Access / CC BY 4.0]

Extends issue resolution to multiple programming languages; the original Python-centric SWE-bench provides a cleaner pedagogical anchor.

04 120 分钟

复习:补救台账收尾与诚实分数更新

独立掌握检查(无 AI 辅助)

前置要求

第 89 天活动 1-3 完成。

学习目标

闭环:用纠正后的前后结果更新掌握台账,让仍未达项保持可见未达,并把辅助证据记入永不并入独立分数的单独列。

操作指引

1. 合并评分表:每个域按第 88 天表与今日复测写出前后小计。
2. 仍 <2 分的项连同具名恢复路径结转到第 90 天后续学习计划——保持可见未达。
3. 活动 3 的判定记入标注 Agent 辅助工程检查的 assisted_evidence 列;它证明委派能力,绝非无辅助知识。
4. 写最终证据说明:哪些工件经命令验证(独立),哪些是自我报告或辅助;列出产生验证的准确命令。
5. 输出 mastery_ledger.json:各域前后、带恢复路径的剩余缺口、independent_evidence 清单、assisted_evidence 清单、诚实范围声明。

自主演练与记录

产出填写完整的 mastery_ledger.json;辅助结果只能出现在 assisted_evidence 下。

独立掌握标准

最终台账诚实分离独立与辅助证据,未达标准保持可见未达,范围声明属实。

权威源资源与权利证明

linking Journal publisher copyright (all rights reserved) - lawful external linking only

Test-Enhanced Learning: Taking Memory Tests Improves Long-Term Retention

作者 / 机构: Henry L. Roediger III, Jeffrey D. Karpicke | 原始出处链接 ↗

对应章节 / 时间戳: Psychological Science 17(3): 249-255, Experiments 1-2

验证版本与日期: Psychological Science, 2006 (Verified: 2026-10-30)

入选理由:

Foundational experimental evidence that retrieval practice (testing) outperforms repeated study for long-term retention; grounds the Day 88 AI-free retrieval checks in peer-reviewed learning science rather than folk advice.

来源前置条件:

Basic understanding of study sessions, retention intervals, and experimental design vocabulary.

安全回退方案(离线弹性):

The curriculum restates the experimental design and effect sizes in original prose; every sealed task runs without the paper.

对比候选资源与选型分析 (2)

Strong meta-analytic coverage of spacing effects but addresses scheduling rather than the retrieval effect itself; retained as support for the Day 90 spaced-review schedule.

Equally foundational follow-up demonstrating repeated retrieval vs repeated study; the 2006 paper is cited as the standard first controlled demonstration.

linking CC BY 4.0

The Turing Way: A Handbook for Reproducible, Ethical and Collaborative Data Science

作者 / 机构: The Turing Way Community | 原始出处链接 ↗

对应章节 / 时间戳: Reproducible Research chapter: version control, testing, environment capture

验证版本与日期: Zenodo DOI 10.5281/zenodo.3233853, continuously updated (Verified: 2026-10-30)

入选理由:

Community-maintained CC-BY handbook codifying reproducibility checklists (versioned code, environment capture, rerunnable commands); directly defines the Day 90 reproduction-verification and provenance-audit workflow.

来源前置条件:

Basic version control and command-line familiarity.

安全回退方案(离线弹性):

The portfolio checklist and verification log schema are authored locally; reproduction commands run without the handbook.

对比候选资源与选型分析 (2)

Concise practices paper; The Turing Way's maintained, checklist-driven coverage of environments and reproducibility is broader and kept current.

Foundational rules list; The Turing Way expands the same rules into actionable checklists used verbatim for the Day 90 verification run.