feat: checkpoint claim codes, badge contracts, and progress verifier (FP-411/412/415)

- verify.py gains a progress subcommand per course: per-checkpoint status
  derived from the objective suite (l03 starter-suite, l04 both-suites)
  plus learner-attested checkpoints (l01/l02/l05), each printing a
  deterministic 8-char base32 claim code derived from documented
  per-course constants (FP-411)
- COURSE.md/COURSE_cn.md gain the badge contract section (badge id, name,
  points, honest self-reported-evidence framing), content_version 2,
  manifest regenerated (FP-412)
- verify_courses.py validates the progress contract: double-run
  determinism, document shape, five checkpoints, claim-code format
  (FP-415)

Author: xxg1413

content-manifest.json

@@ -50,7 +50,7 @@
     {
       "id": "course-agent-rules",
       "type": "course",
-      "content_version": 1,
+      "content_version": 2,
       "status": "reviewed",
       "reviewed_on": "2026-09-12",
       "locales": [
@@ -59,14 +59,14 @@
           "path": "courses/agent-rules-single-source/COURSE.md",
           "title": "One source of truth for agent rules",
           "summary": "Stop maintaining diverging AGENTS.md, CLAUDE.md, and .cursorrules files — build a checker that proves your repository has exactly one rule source, taught hands-on with your coding agent.",
-          "sha256": "41aaa6f0681ea5a40263dd5372f0d4bf402c947b6727486c4ff5ba7f0a2a9cee"
+          "sha256": "b09497cd76bea0191c3944dec61cc9bb89f245a21750fcf5013f4c4d4db7d51c"
         },
         {
           "lang": "zh-CN",
           "path": "courses/agent-rules-single-source/COURSE_cn.md",
           "title": "Agent 规则的单一真源",
           "summary": "不再同时维护彼此打架的 AGENTS.md、CLAUDE.md 与 .cursorrules——亲手构建一个能证明仓库只有一处规则真源的检查器，由你的编码 Agent 授课。",
-          "sha256": "5b90d5556cdff097812ef25ab9134b3939f677f6106c162c28ea15f1986091ee"
+          "sha256": "245fc0cc40f3103ed5fed99f6f656e9b04ff7093bd0d0782203a0dbafa647864"
         }
       ]
     },
@@ -188,7 +188,7 @@
     {
       "id": "course-claude-code",
       "type": "course",
-      "content_version": 1,
+      "content_version": 2,
       "status": "reviewed",
       "reviewed_on": "2026-09-12",
       "locales": [
@@ -197,14 +197,14 @@
           "path": "courses/hands-on-python-with-claude-code/COURSE.md",
           "title": "Hands-on Python with Claude Code",
           "summary": "An agent-taught course that takes you from a cloned folder to a tested, verified Python report tool using Claude Code — including the task contract, the bounded change, and the objective pass/fail evidence.",
-          "sha256": "048b8f210dedc9d60bffb98a8347c89586295e5be3c16abd6be20ea50ef48563"
+          "sha256": "aec892c18cd7446b9d08c69b116d8c00a0bffc6b6ea6d715dfc969298708d5a3"
         },
         {
           "lang": "zh-CN",
           "path": "courses/hands-on-python-with-claude-code/COURSE_cn.md",
           "title": "用 Claude Code 实战 Python",
           "summary": "一门由 Agent 授课的课程：从下载课程文件夹到交付一个经过测试与验证的 Python 报表工具——包括任务契约、最小变更和客观的通过/失败证据。",
-          "sha256": "96c3287051153c42678be558538b9be16cc9835296fc8947711e48ea6ee0a145"
+          "sha256": "185e21a796cac870969f2fff73b3f103eb4a43e215c11be9f0044dc3ec0e7e22"
         }
       ]
     },
@@ -326,7 +326,7 @@
     {
       "id": "course-codex-cli",
       "type": "course",
-      "content_version": 1,
+      "content_version": 2,
       "status": "reviewed",
       "reviewed_on": "2026-09-12",
       "locales": [
@@ -335,14 +335,14 @@
           "path": "courses/hands-on-with-openai-codex-cli/COURSE.md",
           "title": "Hands-on with OpenAI Codex CLI",
           "summary": "The same verified Python workflow as the Claude Code course — task contract, bounded change, objective verify.py evidence — taught hands-on with the OpenAI Codex CLI instead.",
-          "sha256": "aa223ff4356ab25c32f55344362a5133fcba0f25851e05c6bd3c6bde9189a8aa"
+          "sha256": "a0dfd154f544fee2eedfc5a34546286af344afd7e6901abef6dec9ba0967fb76"
         },
         {
           "lang": "zh-CN",
           "path": "courses/hands-on-with-openai-codex-cli/COURSE_cn.md",
           "title": "OpenAI Codex CLI 实战",
           "summary": "与 Claude Code 课程相同的已验证 Python 工作流——任务契约、最小变更、客观 verify.py 证据——改用 OpenAI Codex CLI 完成实战。",
-          "sha256": "4a31ae97c205ee1e28008f1149aadf1488d4d305176678da1c475475174944d5"
+          "sha256": "058985bd5e3f274df28de5ab12ba400eb3ea1367be7daee25a3b98424d4656b3"
         }
       ]
     },
@@ -464,7 +464,7 @@
     {
       "id": "course-mcp-tools",
       "type": "course",
-      "content_version": 1,
+      "content_version": 2,
       "status": "reviewed",
       "reviewed_on": "2026-09-12",
       "locales": [
@@ -473,14 +473,14 @@
           "path": "courses/mcp-server-in-python/COURSE.md",
           "title": "Give your agent tools with MCP (Python)",
           "summary": "Build a stateless Model Context Protocol tool server in pure Python — JSON-RPC 2.0 dispatch, schema validation, error isolation, and the 2026-07-28 input_required round-trip — and learn where a real tool belongs.",
-          "sha256": "bab5f79922a15ac52c736a549bc69a392718fff274e20cae0ad7c3428e00c4f6"
+          "sha256": "7cf318f04fd92925789d11cff48afc249fbe4f96b02e68bfa30fe054f0b6da22"
         },
         {
           "lang": "zh-CN",
           "path": "courses/mcp-server-in-python/COURSE_cn.md",
           "title": "用 MCP 给你的 Agent 装上工具（Python）",
           "summary": "用纯 Python 构建一个无状态的 Model Context Protocol 工具服务——JSON-RPC 2.0 分发、schema 校验、错误隔离与 2026-07-28 的 input_required 多轮交互——并理解真正的工具该放在哪里。",
-          "sha256": "a11236fcd046e9e84e54b22ebcce7e01b925ef71f12014ebb1f1b8bcb536cb63"
+          "sha256": "74af520640128e8df445f5e7af6f8391eb525af26698871b560fd98a6816961e"
         }
       ]
     },
@@ -602,7 +602,7 @@
     {
       "id": "course-verify-ship",
       "type": "course",
-      "content_version": 1,
+      "content_version": 2,
       "status": "reviewed",
       "reviewed_on": "2026-09-12",
       "locales": [
@@ -611,14 +611,14 @@
           "path": "courses/verifying-ai-generated-code/COURSE.md",
           "title": "From \"it runs\" to \"it ships\": verifying AI-generated code",
           "summary": "Build the release-evidence machine your AI-written project is missing — a ship check that runs the tests, parses what actually ran, refuses zero-test projects, and writes an honest delivery record.",
-          "sha256": "7765ea8ad032cc16eb5d34b0f65401c28c8c56cc1f44c496c9cc298f44cb7cd3"
+          "sha256": "d4f2b0448f3c637acfdff97e76ab72e622833f4904376d0fad4e54dbe189ef23"
         },
         {
           "lang": "zh-CN",
           "path": "courses/verifying-ai-generated-code/COURSE_cn.md",
           "title": "从「能跑」到「能上线」：验证 AI 写的代码",
           "summary": "为你的 AI 项目补上缺失的发布证据机器——一个 ship check：运行测试、解析实际跑了什么、拒绝零测试项目，并写出诚实的交付记录。",
-          "sha256": "55eda6add807fff2147c7829da190e8c7cdfdbb537d3dcfdb85b461e2c0a21f2"
+          "sha256": "50b2624cdae9a7597b841ec8f70339d928b431ceff4638e9c3e1e67d90a06188"
         }
       ]
     },

courses/agent-rules-single-source/COURSE.md

@@ -4,7 +4,7 @@ type: course
 title: One source of truth for agent rules
 summary: Stop maintaining diverging AGENTS.md, CLAUDE.md, and .cursorrules files — build a checker that proves your repository has exactly one rule source, taught hands-on with your coding agent.
 lang: en-US
-content_version: 1
+content_version: 2
 status: reviewed
 reviewed_on: 2026-09-12
 ---
@@ -58,8 +58,17 @@ and the AGENTS.md guide on flypython.com), multi-repo setups, or
 machine-policy enforcement. The checker is deliberately narrow: one
 directory, three file names, one truth.
 
+
+## Badge contract
+
+- Badge: **Measure the drift Badge** (badge id `course-agent-rules`) - earned by claiming all five checkpoints.
+- Challenges: L01-L05 checkpoints, 10 points each; +50 course-badge bonus when all five are claimed on flypython.com.
+- Evidence: `python verify.py progress` - L03 (bounded change) and L04 (verify & review) are objectively gated by the suite; L01/L02/L05 are learner-attested.
+- Submission: each passed checkpoint prints a deterministic claim code; record it on flypython.com against your account. Self-reported evidence, never a certificate.
+
 ## Folder map
 
+
 `COURSE.md`/`COURSE_cn.md`, bilingual `lessons/`, `scenario/` repositories,
 `TASK.md`/`TASK_cn.md` (the code contract), `starter/`, `solution/`, `tests/`
 (11 tests), `verify.py`, `REVIEW.md`.

courses/agent-rules-single-source/COURSE_cn.md

@@ -4,7 +4,7 @@ type: course
 title: Agent 规则的单一真源
 summary: 不再同时维护彼此打架的 AGENTS.md、CLAUDE.md 与 .cursorrules——亲手构建一个能证明仓库只有一处规则真源的检查器，由你的编码 Agent 授课。
 lang: zh-CN
-content_version: 1
+content_version: 2
 status: reviewed
 reviewed_on: 2026-09-12
 ---
@@ -53,8 +53,17 @@ reviewed_on: 2026-09-12
 flypython.com 上的 AGENTS.md 指南）、多仓库配置、机器策略强制。检查器
 刻意收窄：一个目录、三个文件名、一个真源。
 
+
+## 徽章契约
+
+- 徽章：**量化漂移徽章**（徽章 id `course-agent-rules`）——认领全部五个检查点后获得。
+- 挑战：L01–L05 检查点各 10 分；五项全部在 flypython.com 认领后另加 50 分课程徽章奖励。
+- 证据：`python verify.py progress` —— L03（边界修改）与 L04（验证与审查）由测试套件客观判定；L01/L02/L05 为学习者自报。
+- 提交：每个通过的检查点会打印确定性认领码，在 flypython.com 上记入你的账号。这是自我报告的证据，绝不是证书。
+
 ## 文件夹结构
 
+
 `COURSE.md`/`COURSE_cn.md`、双语 `lessons/`、`scenario/` 仓库、
 `TASK.md`/`TASK_cn.md`（代码契约）、`starter/`、`solution/`、`tests/`
 （11 个测试）、`verify.py`、`REVIEW.md`。

courses/agent-rules-single-source/verify.py

@@ -7,6 +7,9 @@
 from __future__ import annotations
 
 import argparse
+import base64
+import hashlib
+import json
 import os
 import subprocess
 import sys
@@ -24,12 +27,85 @@
 )
 
 
+
+COURSE_ID = 'course-agent-rules'
+# Documented constant: claim codes derive deterministically from
+# (COURSE_ID, checkpoint_id, COURSE_SALT). They are spot-checkable
+# self-reported evidence, not tamper-proof secrets — see
+# docs/repo-plan-0.0.4.md FP-411.
+COURSE_SALT = '3af23393e1911a4a'
+
+CHECKPOINTS = [
+    {"id": "l01", "gate": "attest", "title": '量化漂移 / Measure the drift'},
+    {"id": "l02", "gate": "attest", "title": '确定唯一真源 / Decide the source of truth'},
+    {"id": "l03", "gate": "starter-suite", "title": '有边界的变更 / Bounded change'},
+    {"id": "l04", "gate": "both-suites", "title": '验证并接入流程 / Verify and wire it in'},
+    {"id": "l05", "gate": "attest", "title": '应用到你的仓库 / Apply to your repository'},
+]
+
+def _claim_code(checkpoint_id):
+    digest = hashlib.sha256(
+        (COURSE_ID + ":" + checkpoint_id + ":" + COURSE_SALT).encode("utf-8")
+    ).digest()
+    return base64.b32encode(digest).decode("ascii")[:8]
+
+def _run_suite(implementation):
+    environment = os.environ.copy()
+    environment["PYTHONPATH"] = str(ROOT / implementation)
+    return subprocess.run(
+        [sys.executable, "-m", "unittest", "discover", "-s", str(ROOT / "tests")],
+        env=environment, check=False, capture_output=True, text=True,
+    )
+
+def run_progress(as_json):
+    starter = _run_suite("starter")
+    solution = _run_suite("solution")
+    starter_ok = starter.returncode == 0
+    solution_ok = solution.returncode == 0
+    rows = []
+    for checkpoint in CHECKPOINTS:
+        gate = checkpoint["gate"]
+        if gate == "attest":
+            status, kind = "attest", "attested"
+        elif gate == "starter-suite":
+            status = "passed" if starter_ok else "open"
+            kind = "objective"
+        else:
+            status = "passed" if (starter_ok and solution_ok) else "open"
+            kind = "objective"
+        code = _claim_code(checkpoint["id"]) if status in ("passed", "attest") else None
+        row = dict(checkpoint)
+        row["status"] = status
+        row["kind"] = kind
+        row["claim_code"] = code
+        rows.append(row)
+    if as_json:
+        print(json.dumps({"course": COURSE_ID,
+            "starter_suite_passed": starter_ok,
+            "solution_suite_passed": solution_ok,
+            "checkpoints": rows}, ensure_ascii=False, indent=2))
+    else:
+        starter_state = "通过" if starter_ok else "未通过"
+        solution_state = "通过" if solution_ok else "未通过"
+        print("课程 " + COURSE_ID)
+        print("实现状态: starter " + starter_state + " / solution " + solution_state)
+        for row in rows:
+            state = row["status"] + ("（自报）" if row["kind"] == "attested" else "")
+            code = "认领码 " + row["claim_code"] if row["claim_code"] else "—"
+            print("  " + row["id"] + "  " + row["title"] + "  [" + state + "]  " + code)
+        print("认领码是自我报告的证据，在 flypython.com 记录；绝非证书。")
+    return 0
+
 def main() -> int:
     parser = argparse.ArgumentParser(description=__doc__)
-    parser.add_argument("implementation", choices=("starter", "solution"))
+    parser.add_argument("implementation", choices=("progress", "starter", "solution"))
+    parser.add_argument("--json", action="store_true")
     parser.add_argument("--expect-failure", action="store_true")
     args = parser.parse_args()
 
+    if args.implementation == "progress":
+        return run_progress(args.json)
+
     command = [sys.executable, "-m", "unittest", "discover", "-s", str(ROOT / "tests")]
     environment = os.environ.copy()
     environment["PYTHONPATH"] = str(ROOT / args.implementation)

courses/hands-on-python-with-claude-code/COURSE.md

@@ -4,7 +4,7 @@ type: course
 title: Hands-on Python with Claude Code
 summary: An agent-taught course that takes you from a cloned folder to a tested, verified Python report tool using Claude Code — including the task contract, the bounded change, and the objective pass/fail evidence.
 lang: en-US
-content_version: 1
+content_version: 2
 status: reviewed
 reviewed_on: 2026-09-12
 ---
@@ -70,8 +70,17 @@ Installing Claude Code, prompt tricks, multi-file architecture, pandas or
 real Excel (`.xlsx`) files, deployment, or selling anything. For those, the
 companion repository has guides and playbooks — link, don't improvise.
 
+
+## Badge contract
+
+- Badge: **Reproduce the failure Badge** (badge id `course-claude-code`) - earned by claiming all five checkpoints.
+- Challenges: L01-L05 checkpoints, 10 points each; +50 course-badge bonus when all five are claimed on flypython.com.
+- Evidence: `python verify.py progress` - L03 (bounded change) and L04 (verify & review) are objectively gated by the suite; L01/L02/L05 are learner-attested.
+- Submission: each passed checkpoint prints a deterministic claim code; record it on flypython.com against your account. Self-reported evidence, never a certificate.
+
 ## Folder map
 
+
 ```
 COURSE.md / COURSE_cn.md   this file (EN / 中文)
 lessons/L01.md … L05.md    lessons (each has an _cn.md pair)

courses/hands-on-python-with-claude-code/COURSE_cn.md

@@ -4,7 +4,7 @@ type: course
 title: 用 Claude Code 实战 Python
 summary: 一门由 Agent 授课的课程：从下载课程文件夹到交付一个经过测试与验证的 Python 报表工具——包括任务契约、最小变更和客观的通过/失败证据。
 lang: zh-CN
-content_version: 1
+content_version: 2
 status: reviewed
 reviewed_on: 2026-09-12
 ---
@@ -62,8 +62,17 @@ Claude Code 安装、提示词技巧、多文件架构、pandas 或真正的 Exc
 （`.xlsx`）文件、部署、任何销售内容。这些主题在配套仓库里有指南和
 手册——链接过去，不要现场发挥。
 
+
+## 徽章契约
+
+- 徽章：**复现故障徽章**（徽章 id `course-claude-code`）——认领全部五个检查点后获得。
+- 挑战：L01–L05 检查点各 10 分；五项全部在 flypython.com 认领后另加 50 分课程徽章奖励。
+- 证据：`python verify.py progress` —— L03（边界修改）与 L04（验证与审查）由测试套件客观判定；L01/L02/L05 为学习者自报。
+- 提交：每个通过的检查点会打印确定性认领码，在 flypython.com 上记入你的账号。这是自我报告的证据，绝不是证书。
+
 ## 文件夹结构
 
+
 ```
 COURSE.md / COURSE_cn.md   本文件（英文 / 中文）
 lessons/L01.md … L05.md    课程（每课都有 _cn.md 中文版）