@tea-agent/loop-agent 0.1.0-fe-test.0

package/docs/templates/backend-test-dag.retrospect.prompt.md

@@ -0,0 +1,165 @@
+# Backend Test DAG Retrospect Prompt Template
+
+## Purpose
+
+Use this prompt for a **test retrospective** node: `executor: "pi"`, `role: "closeout"`, `toolProfile: "write"`, `writePolicy: "exclusive"`. The closeout agent reads **Backend Test Result v1** (and classification), then generates a retrospective report with an objective maturity rating.
+
+Do **not** create a new executor type. This is a standard `executor: pi` writer node.
+
+## Recommended DAG Node Shape
+
+```json
+{
+  "id": "test-retrospect-pi",
+  "depends_on": ["classify-backend-test-result-pi"],
+  "complexity": "MED",
+  "executor": "pi",
+  "role": "closeout",
+  "toolProfile": "write",
+  "writePolicy": "exclusive",
+  "writeSet": ["docs/test-reports/**"],
+  "allowedPaths": ["docs/test-reports/**"],
+  "forbiddenPaths": [".harness/**", "artifacts/**"],
+  "outputContract": "Markdown retrospective report under docs/test-reports/ with coverage summary, review findings, Result v1 stats, classification, and maturity rating (A/B/C/D).",
+  "subtask_prompt_markdown": "./backend-test-dag.retrospect.prompt.md"
+}
+```
+
+## Prompt Body
+
+You are the Backend Test DAG **test retrospective** agent.
+
+Your job is to read upstream Result v1 + classification (+ review report) and generate a retrospective report with a maturity rating. Write the report under `docs/test-reports/` only. Stay within `writeSet`. Do not write root `artifacts/**`.
+
+This node runs on **both pass and assertion-fail** paths after canonical context. Treat Result v1 as the sole testOutcome authority; pipeline completion and L-5 readiness are separate conclusions. **Never** rewrite a failed result as passed in this report. The context also carries advisory case review and deterministic traceability evidence, which must be reflected in the retrospective.
+
+### Output Steps (do in order)
+
+1. First, output the maturity rating on the first line: `Rating: A/B/C/D`
+2. Then write the full report under `docs/test-reports/`
+
+### Inputs
+
+1. **Result v1 (authoritative stats)** — `$HARNESS_DAG_RUN_DIR/contracts/backend-test-result.json`
+   Use `passed` / `failed` / `error` / `skipped` / `outcome` / `failures[]` / `pytestExitCode` only from this artifact.
+2. **Case Manifest v1 (authoritative AC coverage)** — `$HARNESS_DAG_RUN_DIR/contracts/backend-test-case-manifest.json`
+   Use `coverageSummary.acCoverageRatio`, `coveredAcCount`, `explicitAcCount`, case counts only from this artifact.
+3. **Classification** — `classify-backend-test-result-pi` JSON (`category`, `confidence`, `evidence`). Interpretive only; does not override outcome.
+4. **Review report** — `review-backend-cases-pi` output (VERDICT, findings, coverage assessment).
+5. **Code Coverage v1** — optional validated `contracts/code-coverage-v1.json`, generated by coverage.py/pytest-cov or JaCoCo. Never infer it from Result v1.
+7. Optional secondary: execute stdout markers / JUnit path (do not re-parse logs for counts when Result v1 exists).
+
+Do NOT re-read source documents. Use upstream outputs only.
+
+### Stats authority
+
+- Pass rate = `passed / (passed + failed + error)` when denominator > 0 (skipped excluded from denominator unless Result documents otherwise) — **Result v1 only**.
+- AC coverage = `coverageSummary.acCoverageRatio` from Case Manifest v1 only (do **not** recompute or invent percentages).
+- Automation coverage = `coverageSummary.generatedCount / coverageSummary.caseCount`; if either field is missing, output `unavailable`.
+- Code coverage = Code Coverage v1 only. Display line (gate ≥80%), branch (gate ≥70%), and function/method (display only), each with covered, total, ratio, status, reason, source scope, requirement IDs, tool/version, commit and artifact SHA-256. Missing artifacts or line/branch metrics are `unavailable` and block L-5.
+- Failed case table rows must match `failures[]` from Result v1.
+- If Result v1 `outcome` is not `passed`, the retrospective **must not** claim overall success.
+
+### Maturity Rating Criteria
+
+| Rating | Coverage (manifest) | Pass Rate (Result v1) | Review Findings |
+|--------|---------------------|------------------------|-----------------|
+| **A** | `acCoverageRatio` = 1 | 100% pytest pass | No Critical or Important findings |
+| **B** | `acCoverageRatio` ≥ 0.8 | ≥90% pytest pass | Only Informational findings |
+| **C** | `acCoverageRatio` ≥ 0.6 | ≥70% pytest pass | No Critical findings (Important allowed) |
+| **D** | Below C thresholds | Below C thresholds | Or any Critical finding unresolved |
+
+#### Rating Rules
+
+- **Coverage** from Case Manifest `coverageSummary` only (deterministic gate product).
+- **Pass rate** from Result v1 only (not guessed from logs).
+- If `review-backend-cases-pi` returned `VERDICT: request-revision` and revision was not completed, cap at **D**.
+- If Result v1 shows >30% failed+error among executed tests, cap at **D** regardless of coverage.
+- Skipped tests count as "not covered" for pass rate but not as failures.
+- Collection/command/report errors → cap at **D** and record classification (not ProductBug by default).
+- The downstream `l5-metrics-pi` node owns the separate L-5 conclusion. This retrospective must provide the authoritative evidence and risks, but must not replace the downstream L-5 JSON conclusion.
+
+### Report Structure
+
+Write the report as a Markdown file named `backend-test-retrospect-<date>.md` under `docs/test-reports/`.
+
+```markdown
+# Backend Test Retrospective Report
+
+**Date:** <YYYY-MM-DD>
+**Task:** <task-id>
+**Maturity Rating:** <A|B|C|D>
+**Result outcome:** <from Result v1>
+**Classification:** <from classify JSON>
+
+## 1. Test Coverage Summary
+
+| Metric | Value |
+|--------|-------|
+| Total acceptance criteria | N |
+| Covered by test cases | N (X%) |
+| Total functional test cases | N |
+| Automated cases / total cases | N / N (X%) or unavailable |
+| Code line coverage | N / N (X%), threshold ≥80%, status |
+| Code branch coverage | N / N (X%), threshold ≥70%, status |
+| Code function/method coverage | N / N (X%) or unavailable |
+
+Code coverage must include language, tool/version, requirement IDs, source scope, commit and artifact SHA-256.
+
+The downstream `l5-metrics-pi` node consumes this report and the run-owned contracts to calculate the independent L-5 conclusion.
+
+## 2. Automation Results (from Result v1)
+
+| Metric | Value |
+|--------|-------|
+| Total tests | N |
+| Passed | N |
+| Failed | N |
+| Error | N |
+| Skipped | N |
+| Pass rate | X% |
+| Pytest exit code | N |
+| Outcome | … |
+| Execution status | … |
+
+### Failed Test Analysis
+
+| Test Case / Function | Kind | Classification | Confidence | Evidence | Owner direction |
+|----------------------|------|----------------|------------|----------|----------------|
+| ... | ... | ... | ... | ... | ... |
+
+## 3. Defects
+
+Use the local Bug ledger in the same directory as the report. Include defect ID, title, AC/case, severity, priority, status, reproduction, evidence, category, owner, fix version and regression status. If absent, write `缺陷登记：unavailable`.
+
+## 4. Risks
+
+| Risk ID | Level | Impact | Likelihood | Current control | Residual risk | Treatment |
+|---------|-------|--------|------------|-----------------|---------------|-----------|
+| ... | Critical/High/Medium/Low | ... | ... | ... | ... | ... |
+
+## 5. Regression Recommendations
+
+Every recommendation must reference a failure/risk/AC/case ID and include target suite, priority and verification command.
+
+## 6. Review Findings
+
+| Severity | Finding | Status |
+|----------|---------|--------|
+| … | … | … |
+
+## 7. Maturity Rating Rationale
+
+Explain which threshold was met or missed.
+
+## 8. Recommendations
+
+- Actionable items for the next iteration.
+- Do not propose changing production code solely to greenwash tests.
+```
+
+### Output Shape (after rating line)
+
+After the mandatory maturity rating line, provide a brief summary paragraph before writing the full report file.
+
+Do not include chain-of-thought. Do not write root `artifacts/**`.

package/docs/templates/backend-test-dag.review-cases.prompt.md

@@ -0,0 +1,36 @@
+# Backend Test DAG Review Cases Prompt Template
+
+## Purpose
+
+供 `review-and-revise-backend-md-cases-pi` 使用。该节点是 `executor: "pi"`、`role: "reviewer"`、`toolProfile: "write"` 的受限 reviewer，只能直接修订 `testcase/md/**`，不得生成 pytest 或修改任务源、生产代码和配置。
+
+## 当前合同
+
+把 Markdown 用例视为面向测试、研发、产品与评审人员的正式文档，而不是模型中间产物。
+
+### 阅读体验
+
+- `testcase/md/README.md` 是简洁入口，包含测试目标、环境、隔离/清理策略、模块汇总和可跳转的用例索引。
+- 模块文件采用中文用例卡片；每条以 `## BE-<MODULE>-<NNN>｜<中文用例名称>` 开始；`<NNN>` 必须是三位补零序号（`001`，禁止 `01`）。Reviewer 发现两位序号或模块下划线时，必须在标题、README 索引与自动化映射中一致规范化（如 `BE-RESOURCE_NOTES-01` → `BE-RESOURCE-NOTES-001`）。
+- 新文档优先使用：`测试目的`、`验收标准`、`需求依据`、`前置条件`、`测试数据`、`操作步骤`、`预期结果`、`自动化映射`。
+- validator 同时接受上述中文分节和历史英文分节；机器 ID、HTTP 方法、路径、字段、枚举、文件名、函数名与 source citation 必须保持精确。
+- 步骤和预期可以用紧凑表格，也可以分别使用编号/项目列表；必须可执行、可独立断言。
+- 自动化内部限制应简短或放进 `<details>`，不能淹没人类主要阅读路径。
+
+### 正确性
+
+- 对照每条 `需求依据` 和环境报告检查 AC、接口、字段/响应形状、状态码、错误语义、状态转换、正向/异常/边界场景。
+- 删除无依据场景、合并重复用例、补齐有依据的遗漏；无法确认的内容写入中文证据缺口，不猜测行为或凭据。
+- 拒绝“符合预期”“正常工作”等模糊结果，以及无意义的中英双写和大段重复 boilerplate。
+- 若能确定脚本与函数命名，在 `自动化映射` 中写明计划脚本路径与 pytest 函数/方法名；脚本路径必须等于模块 one-to-one 路径 `testcase/test_<module>.py`，`<module>` 来自当前 Markdown 文件名 stem，不得使用 Case-ID 式模块名（如 `BE-HEALTH.md`）或与模块 stem 不一致的映射。
+
+## 推荐输出
+
+完成文件修订后，仅用简短中文说明：
+
+- 修订了哪些文档；
+- 用例数与主要模块；
+- 修复了哪些需求一致性或阅读问题；
+- 仍有哪些证据缺口。
+
+不得输出 JSON，不得执行测试。

package/docs/templates/backend-test-execution.schema.json

@@ -0,0 +1,138 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "https://tea-agent.dev/schemas/backend-test-execution-v1.json",
+  "title": "Backend Test Execution v1",
+  "type": "object",
+  "additionalProperties": false,
+  "required": [
+    "schemaVersion",
+    "framework",
+    "runner",
+    "testRoot",
+    "workingDirectory",
+    "report",
+    "targetMode",
+    "existingFixtures",
+    "authenticationMode",
+    "requiredEnvNames",
+    "dataIsolation",
+    "evidenceGaps",
+    "evidenceRefs"
+  ],
+  "properties": {
+    "schemaVersion": { "const": 1 },
+    "framework": { "const": "pytest" },
+    "runner": {
+      "type": "object",
+      "additionalProperties": false,
+      "properties": {
+        "commandParts": {
+          "type": "array",
+          "items": { "type": "string", "minLength": 1 }
+        },
+        "frozenCommandHints": {
+          "type": "array",
+          "items": { "type": "string", "minLength": 1 }
+        }
+      }
+    },
+    "testRoot": {
+      "type": "string",
+      "minLength": 1,
+      "description": "Repo-relative posix path; no absolute form or .. segments"
+    },
+    "workingDirectory": {
+      "type": "string",
+      "minLength": 1,
+      "default": "."
+    },
+    "report": {
+      "type": "object",
+      "additionalProperties": false,
+      "required": ["format", "relativeHint"],
+      "properties": {
+        "format": { "const": "junit" },
+        "relativeHint": { "type": "string", "minLength": 1 }
+      }
+    },
+    "targetMode": {
+      "enum": ["in-process", "external-running-service", "managed-command"]
+    },
+    "baseUrlEnvName": {
+      "type": "string",
+      "pattern": "^[A-Z_][A-Z0-9_]*$"
+    },
+    "readiness": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "additionalProperties": false,
+        "required": ["path", "description"],
+        "properties": {
+          "path": {
+            "type": "string",
+            "minLength": 1,
+            "pattern": "^/(?!/)(?!.*(?:^|/)\\.{1,2}(?:/|$))(?!.*//)(?!.*\\\\)[^?#]*$",
+            "description": "HTTP URL path on the target service for readiness probes (e.g. /api/health). Not a repository file path; must start with one /, use POSIX separators, and must not include duplicate slashes, dot segments, query, fragment, or scheme. Runtime trims surrounding whitespace before validating and materializing the normalized path."
+          },
+          "description": { "type": "string", "minLength": 1 }
+        }
+      }
+    },
+    "existingFixtures": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "additionalProperties": false,
+        "required": ["name", "sourcePath", "kind"],
+        "properties": {
+          "name": { "type": "string", "minLength": 1 },
+          "sourcePath": { "type": "string", "minLength": 1 },
+          "kind": { "type": "string", "minLength": 1 }
+        }
+      }
+    },
+    "authenticationMode": { "type": "string", "minLength": 1 },
+    "requiredEnvNames": {
+      "type": "array",
+      "items": {
+        "type": "string",
+        "pattern": "^[A-Z_][A-Z0-9_]*$"
+      }
+    },
+    "dataIsolation": {
+      "type": "object",
+      "additionalProperties": false,
+      "required": ["mode"],
+      "properties": {
+        "mode": { "type": "string", "minLength": 1 },
+        "evidence": { "type": "string", "minLength": 1 }
+      }
+    },
+    "managedCommand": {
+      "type": "object",
+      "additionalProperties": false,
+      "properties": {
+        "start": { "type": "string", "minLength": 1 },
+        "stop": { "type": "string", "minLength": 1 },
+        "sourceRef": { "type": "string", "minLength": 1 }
+      }
+    },
+    "evidenceGaps": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "additionalProperties": false,
+        "required": ["description"],
+        "properties": {
+          "description": { "type": "string", "minLength": 1 },
+          "sourceRef": { "type": "string", "minLength": 1 }
+        }
+      }
+    },
+    "evidenceRefs": {
+      "type": "array",
+      "items": { "type": "string", "minLength": 1 }
+    }
+  }
+}

package/docs/templates/backend-test-result.schema.json

@@ -0,0 +1,99 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "https://tea-agent.dev/schemas/backend-test-result-v1.json",
+  "title": "Backend Test Result v1",
+  "description": "Run-owned pytest result artifact. Task Pool may consume outcome, counts, failures[], executionStatus, collectionStatus, pytestExitCode, and junit refs. Auto follow-up is out of scope for M2.",
+  "type": "object",
+  "additionalProperties": false,
+  "required": [
+    "schemaVersion",
+    "executionStatus",
+    "pytestExitCode",
+    "collectionStatus",
+    "tests",
+    "passed",
+    "failed",
+    "error",
+    "skipped",
+    "junit",
+    "commandSummary",
+    "failures",
+    "outcome"
+  ],
+  "properties": {
+    "schemaVersion": { "const": 1 },
+    "executionStatus": {
+      "type": "string",
+      "enum": ["completed", "collection-error", "command-error", "report-error"],
+      "description": "Process-level status. Task Pool: do not treat collection-error/command-error as ProductBug."
+    },
+    "pytestExitCode": {
+      "type": "integer",
+      "minimum": 0,
+      "maximum": 255,
+      "description": "Raw pytest process exit code (0/1 are node-success when JUnit exists)."
+    },
+    "collectionStatus": {
+      "type": "string",
+      "enum": ["ok", "error", "unknown", "skipped"]
+    },
+    "tests": {
+      "type": "integer",
+      "minimum": 0,
+      "description": "Task Pool: total tests = passed+failed+error+skipped"
+    },
+    "passed": { "type": "integer", "minimum": 0 },
+    "failed": { "type": "integer", "minimum": 0 },
+    "error": { "type": "integer", "minimum": 0 },
+    "skipped": { "type": "integer", "minimum": 0 },
+    "durationMs": { "type": "number", "minimum": 0 },
+    "junit": {
+      "type": "object",
+      "additionalProperties": false,
+      "required": ["relativePath", "sha256"],
+      "properties": {
+        "relativePath": {
+          "type": "string",
+          "minLength": 1,
+          "pattern": "^(?!/)(?!.*(?:^|/)\\.\\.(?:/|$))(?!.*(?:^|/)\\.(?:/|$))(?!.*\\\\)(?!.*//).+$",
+          "description": "Run-dir relative POSIX path without absolute form, backslashes, dot segments, or parent traversal (e.g. reports/backend-test-junit.xml)"
+        },
+        "sha256": {
+          "type": "string",
+          "pattern": "^[a-f0-9]{64}$"
+        }
+      }
+    },
+    "commandSummary": {
+      "type": "string",
+      "minLength": 1,
+      "description": "Non-secret command summary only"
+    },
+    "failures": {
+      "type": "array",
+      "description": "Truncated failure/error summaries for triage (Task Pool consumable).",
+      "items": {
+        "type": "object",
+        "additionalProperties": false,
+        "required": ["classname", "name", "message"],
+        "properties": {
+          "classname": { "type": "string", "minLength": 1 },
+          "name": { "type": "string", "minLength": 1 },
+          "message": { "type": "string", "minLength": 1 },
+          "kind": { "type": "string", "enum": ["failure", "error"], "default": "failure" }
+        }
+      }
+    },
+    "outcome": {
+      "type": "string",
+      "enum": [
+        "passed",
+        "completed-with-failures",
+        "collection-error",
+        "command-error",
+        "report-error"
+      ],
+      "description": "Authoritative shell-facing outcome for backend-test-outcome-gate-shell. Retrospective must not override."
+    }
+  }
+}

package/docs/templates/branch-merge-report.md

@@ -0,0 +1,115 @@
+# Branch Merge Report — `<source-ref>` into `<target-branch>` at `<source-short-sha>`
+
+Date: `<YYYY-MM-DD>`
+Target branch: `<target-branch>`
+Target before merge: `<full SHA>`
+Source: `<source-ref>@<full SHA>`
+Merge base: `<full SHA>`
+Merge commit: `<full SHA or follow-up commit reference>`
+Merge mode: `quick | standard | deep`
+Mode rationale: `<why this mode is sufficient>`
+
+## Naming Convention
+
+Use `YYYY-MM-DD-origin-<source>-into-<target>-<source-short-sha>.md` under `docs/reports/merge/`. Replace `/` and other path separators in branch names with `-`. The source short SHA distinguishes repeated merges on the same day.
+
+## Pre-merge Inspection
+
+- Divergence: `<target-only count>` target-only commits / `<source-only count>` source-only commits.
+- Dry run: `git merge-tree --write-tree --messages <target> <source>`.
+- Predicted conflicts: `<count and files>`.
+- Baseline verification: `<commands and results>`.
+
+## Final Source Drift Check
+
+- Final fetch: `<command and timestamp>`.
+- Final source tip: `<full SHA>`.
+- Source advanced during work: `<no | yes, with re-audit details>`.
+
+## Incoming Capabilities
+
+| Capability | Key files / contracts | Expected user or runtime result | Verification |
+| --- | --- | --- | --- |
+| `<incoming capability>` | `<paths>` | `<behavior>` | `<test/smoke>` |
+
+## Target-branch Capabilities to Preserve
+
+| Capability | Key files / contracts | Preservation risk | Verification |
+| --- | --- | --- | --- |
+| `<target capability>` | `<paths>` | `<possible overwrite>` | `<test/smoke>` |
+
+## Conflict Resolution
+
+| File | Conflict type | Incoming concern | Target concern | Resolution | Evidence |
+| --- | --- | --- | --- | --- | --- |
+| `<path>` | `content/rename-delete/add-add` | `<incoming behavior>` | `<target behavior>` | `<combined decision>` | `<test/diff>` |
+
+If there were no textual conflicts, write `No textual conflicts` and continue with the automatic-merge review below.
+
+## Automatic-merge Semantic Review
+
+Record overlapping or high-risk files that Git merged without conflict. These are common places for silent feature loss.
+
+| File / area | Why high risk | Incoming behavior retained | Target behavior retained | Evidence |
+| --- | --- | --- | --- | --- |
+| `<path>` | `<shared imports/topology/config>` | `<incoming proof>` | `<target proof>` | `<tests/search>` |
+
+## Functional Impact Matrix
+
+| Feature | Before merge | After merge | Deleted or weakened? | Notes |
+| --- | --- | --- | --- | --- |
+| `<feature>` | `<baseline>` | `<result>` | `No/Yes` | `<details>` |
+
+## Init and Update Impact
+
+- Changed init/package surface: `<paths or none>`.
+- Fresh full init result: `<result>`.
+- `init check-update` result: `<result>`.
+- `init update --apply-safe` recovery result: `<result>`.
+- Manifest or init implementation updates made: `<details or none>`.
+
+## Package Surface Impact
+
+- `npm pack --dry-run --ignore-scripts --json`: `<file count>` files.
+- Added package paths: `<paths>`.
+- Removed package paths: `<paths or none>`.
+- Unexpected files or missing required files: `<details or none>`.
+
+## Verification
+
+| Command | Result | Capability covered |
+| --- | --- | --- |
+| `npm run typecheck` | `<pass/fail>` | TypeScript integration |
+| `npm test` | `<pass/fail>` | Full regression |
+| `npm run build` | `<pass/fail>` | Build/package inputs |
+| `bash scripts/check-repo.sh` | `<pass/fail>` | Governance and init surface |
+| `npm run docs:build` | `<pass/fail>` | User documentation |
+| `npm pack --dry-run --ignore-scripts --json` | `<pass/fail>` | Published files |
+
+Add focused tests and real init/update/browser/CLI smokes required by the merged capabilities.
+
+### Full-suite authority (local same-tree dedupe)
+
+- Canonical full-verification authority: `<valid receipt | target pre-push | npm run verify:tree | authorized skip/release workflow>`
+- Verified tree SHA: `<git rev-parse HEAD^{tree} or equivalent>`
+- Receipt status/path: `<matched|missing|expired|invalid|ineligible-dirty|n/a; path if any>`
+- Full-suite local execution count: `<number of full local suite runs for this merge tree>`
+- Additional full run reason: `<none | env drift | new merge tree | command contract change | other>`
+- Apparent duplicate skipped and evidence: `<yes/no; receipt matched log or explain output summary>`
+
+## Mode Completion Check
+
+- Required gates for selected mode: `<list>`.
+- Passed gates: `<list>`.
+- Any mode upgrade during execution: `<no | from quick/standard to standard/deep, with reason>`.
+
+## Residual Risk
+
+- `<unverified external environment, performance, migration, or compatibility risk>`
+
+## Rollback and Investigation Pointers
+
+- Merge commit: `<SHA>`.
+- Conflict files: `<paths>`.
+- First tests to run if a feature appears missing: `<commands>`.
+- Safe rollback strategy: prefer `git revert -m 1 <merge-commit>` after review; do not use destructive reset on shared work.

package/docs/templates/evaluation/agents-map-slim-v1.candidate.json

@@ -0,0 +1,9 @@
+{
+  "schemaVersion": 1,
+  "candidateKind": "prompt",
+  "variant": "slim-v1",
+  "bodyPath": "docs/templates/evaluation/agents-map-slim-v1.md",
+  "bodySha256": "0101b0f85d1a78532d9f226f42ca1d3192d3bd421c82310621915b7f00b16968",
+  "lineCount": 87,
+  "purpose": "Eval Lab context-engineering AGENTS map A/B companion body; JSON contentRef avoids private-verifier token leak"
+}

package/docs/templates/evaluation/agents-map-slim-v1.md

@@ -0,0 +1,87 @@
+<!-- CODEGRAPH_START -->
+## CodeGraph
+
+如果仓库根目录存在 `.codegraph/`，在理解或定位代码前优先使用 CodeGraph，再考虑 rg/fd 或手动读文件。
+<!-- CODEGRAPH_END -->
+
+# AGENTS.md
+
+本仓库采用“人类掌舵，智能体执行”的工程方式。目标不是一次性写完所有代码，而是在一个可持续演进、可交接、可验证的系统里做小步增量。
+
+`AGENTS.md` 是地图，不是百科。顶层只保留开工协议、会话协议与文档导航；长期知识、方法论、决策、计划、报告和模板应进入 `docs/`。
+
+## 默认立场
+
+- 仓库是记录系统：决策、契约、计划、测试、报告优先落到仓库，而不是停留在聊天里。
+- 一次只推进一个清晰工作块；主会话按 Orient → Select → Contract → Implement → Verify → Handoff 治理，runtime 真实流程以 `src/workflows/` 为准。
+- 先验证基线，再叠加改动；完成定义必须可验证，不能靠删测试、降标准或模糊描述制造“完成”。
+- Do not consider backward compatibility. Ignore legacy code/libraries.
+- 搜索先于实现；受治理 Agent runtime 只有 Pi（`implement-pi` / `repair-pi`）；`cursor-prompt` 仅为显式手工 one-shot sidecar。
+- DAG 标准路径：Contract → Scout → Plan → Implement → Verify → Closeout/Handoff。
+- 机器校验契约真源：`docs/init-surface.manifest.json`、`docs/architecture/runtime-boundaries.md`、`src/cli/command-definitions.ts`、`skills/loop-agent/`、`scripts/check-*.sh`；本文件只指路。
+- 本仓库既是源项目也是 init 默认模板；新增能力必须判断 npm 内置 vs `loop-agent init` 投影。
+- 委托写入前必须结构化 `task.json.allowedPaths` / `forbiddenPaths`，并审查 DAG writer `writeSet`。
+- Shell 搜索优先 `rg`，按名找文件优先 `fd`；脚本用 Git Bash / 兼容 Bash。
+- 用 loop-agent 迭代本仓库时，控制器必须来自已发布 npm 包（记录实际版本）；启动后不要中途升级；不要用工作区 `npm link` / `npm run dev` 控制可能改 CLI/runtime/package 的任务。
+- 反复出现的约束固化为文档、脚本、检查、测试或模板；禁止占位实现（除非 contract 标明脚手架）。
+- 非微小实现：`exec-plan` 不能替代 Agent DAG；除非用户要求 one-shot 或计划记录 escape hatch，否则改实现前完成 `task advance`（结构化路径 + writeSet 审查），不得主会话直接改实现。
+- 不要自行引入外部 SDD/spec-first 等强制平行治理树；以本文件与 `docs/` 为准（ADR 0006）。
+
+## 开始顺序
+
+改文件前必须先完成：
+
+1. `pwd` → 读 `README.md`、`harness.json`、`docs/README.md`；有 `CONTEXT.md` 则读术语表。
+2. 实现类工作继续读：`docs/governance/development-principles.md`、`docs/governance/feature-workflow.md`、`docs/governance/verification-matrix.md`。
+3. 涉及命令/executor/init/skills/发布包/治理检查时继续读：`docs/architecture/runtime-boundaries.md`、`docs/runtime/loop-agent-harness.md`。
+4. 涉及测试纪律/验证声明/调试时继续读：`docs/harness-methodology-*.md`。
+5. 查看最近提交、相关 plan/progress/report；`git status --short --branch`；跑最小基线验证。
+6. 后端/接口/pytest → `taskKind: "backend-test"`（不是 `--profile`）；知识回写 `knowledge-sync`；图谱开荒 `knowledge-graph-bootstrap`。`--profile` 仅 `auto|minimal|standard|reviewed|supervised`。
+7. 看板/observe → `agent-worker console`（默认 repo=当前目录、port=8790；`--no-open` 禁止自动打开浏览器）（`/inspect/` 只读）；`observe serve` 已下线（REMOVED / exit 2）；`observe snapshot` 仍可用。
+8. 分支合并 → 先读 `docs/operations/branch-merge-guideline.md`。
+
+## 会话协议
+
+1. Orient → 2. Select（一块）→ 3. Contract → 4. Implement → 5. Verify → 6. Converge Docs → 7. Handoff。
+
+这不是 DAG 节点序列。复杂实现默认 Agent DAG；主 agent 拆任务、写 contract、结构化路径、审查 DAG/writeSet/profile/shell verification。微小任务可用 one-shot escape hatch 并记录边界与验证证据。
+
+## 项目地图
+
+- `CONTEXT.md`：术语表
+- `src/`：运行时；`test/`：Vitest；`bin/loop-agent.js`：CLI
+- `skills/`：源仓库/npm 内置 skills；目标项目只生成 `.agents/skills/`
+- `.harness/`：运行态（tasks/dag-runs/runs 等；init 会 gitignore 运行事实，保留 prompts 与占位）
+- `docs/`：治理；`website/`：用户文档站；`scripts/`：检查与 CI
+
+## 工作规则（增量约束）
+
+- 保留无关用户改动；优先沿用现有 helper/目录边界。
+- 长期决策写入 `docs/`；面向用户变更更新 `CHANGELOG.md`（结果导向中文）。
+- init/投影变更必须同步目标项目生成物与 package assets；init evolution 按 `docs/init-surface.manifest.json` 分级。
+- CLI/skill entry/runtime boundary/发布包变更同步 catalog、脚本与测试。
+- 没有新鲜验证证据时不声明完成；新债写入 plan/progress/report。
+
+## 验证
+
+权威源：`docs/governance/verification-matrix.md`。常用：
+
+```bash
+npm run typecheck && npm test && npm run build
+node bin/loop-agent.js --help
+bash scripts/check-repo.sh
+bash scripts/ci.sh
+```
+
+文档站变更：`npm run docs:build`。init / architecture / skill entry / pack 定向验证见 verification-matrix。
+
+## 交接
+
+记录：改了什么、为什么、验证命令与结果、契约/文档/测试影响、剩余风险、后续工作。
+
+## 禁止事项
+
+- 未读相关文档就大改；一次混合无关重构/新功能/文档大迁移。
+- 把对话约束当长期知识；缺验证宣称完成；假设系统没有某能力（先搜索）。
+- stub/假数据通路替代交付；把本机绝对路径写入仓库级 AGENTS/README/模板/发布包。
+- 只更新本仓库体验而遗漏目标项目 init 体验。

package/docs/templates/evaluation/agents-map-verbose-v0.candidate.json

@@ -0,0 +1,9 @@
+{
+  "schemaVersion": 1,
+  "candidateKind": "prompt",
+  "variant": "verbose-v0",
+  "bodyPath": "docs/templates/evaluation/agents-map-verbose-v0.md",
+  "bodySha256": "50af8ea3c7058071f249f9b45bd8c05fb1b1aa257ec6b2667aa91346c106cf81",
+  "lineCount": 153,
+  "purpose": "Eval Lab context-engineering AGENTS map A/B companion body; JSON contentRef avoids private-verifier token leak"
+}