@ranger1/dx 0.1.76 → 0.1.78

package/@opencode/agents/pr-review-aggregate.md

@@ -1,151 +0,0 @@
----
-description: aggregate PR reviews + create fix file
-mode: subagent
-model: openai/gpt-5.3-codex
-temperature: 0.1
-tools:
-  bash: true
----
-
-# PR Review Aggregator
-
-## Cache 约定（强制）
-
-- 缓存目录固定为 `./.cache/`；交接一律传 `./.cache/<file>`（repo 相对路径），禁止 basename-only（如 `foo.md`）。
-
-## 输入（两种模式）
-
-### 模式 A：评审聚合 + 生成 fixFile + 发布评审评论
-
-- `PR #<number>`
-- `round: <number>`
-- `runId: <string>`（必须透传，格式 `<PR>-<ROUND>-<HEAD_SHORT>`，禁止自行生成）
-- `contextFile: <path>`（例如：`./.cache/pr-context-...md`）
-- `reviewFile: <path>`（多行，1+ 条；例如：`./.cache/review-...md`）
-
-### 模式 B：发布修复评论（基于 fixReportFile）
-
-- `PR #<number>`
-- `round: <number>`
-- `runId: <string>`（必须透传，格式 `<PR>-<ROUND>-<HEAD_SHORT>`，禁止自行生成）
-- `fixReportFile: <path>`（例如：`./.cache/fix-report-...md`）
-
-示例：
-
-```text
-PR #123
-round: 1
-runId: 123-1-a1b2c3d
-  contextFile: ./.cache/pr-context-pr123-r1-123-1-a1b2c3d.md
-  reviewFile: ./.cache/review-CDX-pr123-r1-123-1-a1b2c3d.md
-  reviewFile: ./.cache/review-CLD-pr123-r1-123-1-a1b2c3d.md
-  reviewFile: ./.cache/review-GMN-pr123-r1-123-1-a1b2c3d.md
-```
-
-## 执行方式（强制）
-
-所有确定性工作（解析/聚合/发评论/生成 fixFile/输出 JSON）都由 `~/.opencode/agents/pr_review_aggregate.py` 完成。
-
-你只做两件事：
-
-1) 在模式 A 里用大模型判断哪些 finding 是重复的，并把重复分组作为参数传给脚本（不落盘）。
-2) 调用脚本后，把脚本 stdout 的 JSON **原样返回**给调用者（不做解释/分析）。
-
-## 重复分组（仅作为脚本入参）
-
-你需要基于所有 `reviewFile` 内容判断重复 finding 分组，生成**一行 JSON**（不要代码块、不要解释文字、不要换行）。
-
-注意：这行 JSON **不是你的最终输出**，它只用于生成 `--duplicate-groups-b64` 传给脚本。
-
-```json
-{"duplicateGroups":[["CDX-001","CLD-003"],["GMN-002","CLD-005","CDX-004"]]}
-```
-
-## 智能匹配（仅在模式 A + decision-log 存在时）
-
-如果 decision-log（`./.cache/decision-log-pr<PR_NUMBER>.md`）存在，你需要基于 LLM 判断每个新 finding 与已决策问题的本质是否相同，从而生成 **escalation_groups** 参数。
-
-**匹配原则**：
-- **Essence 匹配**：对比 `essence` 字段与新 finding 的问题本质。
-- **文件强绑定**：仅当 decision-log 条目的 `file` 与新 finding 的 `file` **完全一致**时才进行匹配。
-  - 若文件被重命名/删除/拆分，视为不同问题（为了稳定性，不处理复杂的 rename 映射）。
-  - 若 decision-log 条目缺少 `file` 字段（旧数据），则跳过匹配（视为不相关）。
-
-**流程**：
-
-1. 读取 decision-log，提取已 rejected 问题的 `essence` 和 `file` 字段
-2. 逐个新 finding，**先检查 file 是否匹配**
-   - 若 file 不匹配 → 视为 New Issue
-   - 若 file 匹配 → 继续对比 essence
-3. 若 essence 也匹配（"问题本质相同"）：
-4. 收集可升级的问题（重新质疑阈值）：
-   - **升级阈值**：优先级差距 ≥ 2 级
-   - 例如：已 rejected P3 but finding 为 P1 → 可升级质疑
-   - 例如：已 rejected P2 but finding 为 P0 → 可升级质疑
-   - 例如：已 rejected P2 but finding 为 P1 → 不升级（仅差 1 级）
-5. 生成**一行 JSON**（不要代码块、不要解释文字、不要换行），结构如下：
-
-```json
-{"escalationGroups":[["CDX-001"],["GMN-002","CLD-005"]]}
-```
-
-其中每个组表示「可以作为已 rejected 问题的升级质疑」的 finding ID 集合。若无可升级问题，输出空数组：
-
-```json
-{"escalationGroups":[]}
-```
-
-注意：escalation_groups JSON **不是你的最终输出**，它只用于生成 `--escalation-groups-b64` 传给脚本。
-
-## 调用脚本（强制）
-
-模式 A（带 reviewFile + 重复分组 + 智能匹配）：
-
-```bash
-python3 ~/.opencode/agents/pr_review_aggregate.py \
-  --pr <PR_NUMBER> \
-  --round <ROUND> \
-  --run-id <RUN_ID> \
-  --context-file <CONTEXT_FILE> \
-  --review-file <REVIEW_FILE_1> \
-  --review-file <REVIEW_FILE_2> \
-  --review-file <REVIEW_FILE_3> \
-  --duplicate-groups-b64 <BASE64_JSON> \
-  --decision-log-file ./.cache/decision-log-pr<PR_NUMBER>.md \
-  --escalation-groups-b64 <BASE64_JSON>
-```
-
-**参数说明**：
-
-- `--duplicate-groups-b64`：base64 编码的 JSON，格式同上，例如 `eyJkdXBsaWNhdGVHcm91cHMiOltbIkNEWC0wMDEiLCJDTEQtMDAzIl1dfQ==`
-- `--decision-log-file`：decision-log 文件路径（可选；若不存在则跳过智能匹配逻辑）
-- `--escalation-groups-b64`：base64 编码的 escalation groups JSON，格式如上，例如 `eyJlc2NhbGF0aW9uR3JvdXBzIjpbWyJDRFgtMDAxIl1dfQ==`
-
-模式 B（带 fixReportFile）：
-
-```bash
-python3 ~/.opencode/agents/pr_review_aggregate.py \
-  --pr <PR_NUMBER> \
-  --round <ROUND> \
-  --run-id <RUN_ID> \
-  --fix-report-file <FIX_REPORT_FILE>
-```
-
-## 脚本输出处理（强制）
-
-- 脚本 stdout 只会输出**单一一行 JSON**（可 `JSON.parse()`）。
-- **成功时**：你的最终输出必须是**脚本 stdout 的那一行 JSON 原样内容**。
-  - 典型返回：`{"stop":true}` 或 `{"stop":false,"fixFile":"..."}` 或 `{"ok":true}`
-  - 禁止：解释/分析/补充文字
-  - 禁止：代码块（```）
-  - 禁止：前后空行
-- **失败/异常时**：
-  - 若脚本 stdout 已输出合法 JSON（包含 `error` 或其他字段）→ 仍然**原样返回该 JSON**。
-  - 若脚本未输出合法 JSON / 退出异常 → 仅返回一行 JSON：`{"error":"PR_REVIEW_AGGREGATE_AGENT_FAILED"}`（必要时可加 `detail` 字段）。
-
-## fixFile 结构（补充说明）
-
-脚本在模式 A 下生成的 fixFile 分为两段：
-
-- `## IssuesToFix`：只包含 P0/P1（必须修）
-- `## OptionalIssues`：包含 P2/P3（由 pr-fix 自主决定是否修复/或拒绝并说明原因）

package/@opencode/agents/pr_context.py

@@ -1,351 +0,0 @@
-#!/usr/bin/env python3
-# PR context builder (deterministic).
-# - Reads PR metadata + recent comments via gh
-# - Reads changed files via git diff (no patch)
-# - Writes Markdown context file to project cache: ./.cache/
-# - Prints exactly one JSON object to stdout
-
-import argparse
-import json
-import os
-import re
-import subprocess
-import sys
-from urllib.parse import urlparse
-from pathlib import Path
-
-
-def _repo_root():
-    # Prefer git top-level so the cache follows the current repo.
-    try:
-        p = subprocess.run(
-            ["git", "rev-parse", "--show-toplevel"],
-            stdout=subprocess.PIPE,
-            stderr=subprocess.DEVNULL,
-            text=True,
-        )
-        out = (p.stdout or "").strip()
-        if p.returncode == 0 and out:
-            return Path(out)
-    except Exception:
-        pass
-    return Path.cwd()
-
-
-def _cache_dir(repo_root):
-    return (repo_root / ".cache").resolve()
-
-
-def _repo_relpath(repo_root, p):
-    try:
-        rel = p.resolve().relative_to(repo_root.resolve())
-        return "./" + rel.as_posix()
-    except Exception:
-        # Fallback to basename-only.
-        return os.path.basename(str(p))
-
-
-REPO_ROOT = _repo_root()
-CACHE_DIR = _cache_dir(REPO_ROOT)
-MARKER_SUBSTR = "<!-- pr-review-loop-marker"
-
-
-def _json_out(obj):
-    sys.stdout.write(json.dumps(obj, ensure_ascii=True))
-    sys.stdout.write("\n")
-
-
-def _run_capture(cmd):
-    try:
-        p = subprocess.run(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True)
-        return p.returncode, p.stdout, p.stderr
-    except FileNotFoundError as e:
-        return 127, "", str(e)
-
-
-def _detect_git_remote_host():
-    # Best-effort parse from origin remote.
-    rc, origin_url, _ = _run_capture(["git", "remote", "get-url", "origin"])
-    if rc != 0:
-        rc, origin_url, _ = _run_capture(["git", "config", "--get", "remote.origin.url"])
-    if rc != 0:
-        return None
-
-    url = (origin_url or "").strip()
-    if not url:
-        return None
-
-    # Examples:
-    # - git@github.com:owner/repo.git
-    # - ssh://git@github.company.com/owner/repo.git
-    # - https://github.com/owner/repo.git
-    if url.startswith("git@"):  # SCP-like syntax
-        m = re.match(r"^git@([^:]+):", url)
-        return m.group(1) if m else None
-
-    if url.startswith("ssh://") or url.startswith("https://") or url.startswith("http://"):
-        try:
-            parsed = urlparse(url)
-            return parsed.hostname
-        except Exception:
-            return None
-
-    return None
-
-
-def _clip(s, n):
-    if s is None:
-        return ""
-    s = str(s)
-    return s if len(s) <= n else (s[:n] + "...")
-
-
-def _safe_basename(name):
-    if not name:
-        return None
-    base = os.path.basename(name.strip())
-    if base != name.strip():
-        return None
-    if base in (".", ".."):
-        return None
-    return base
-
-
-def _git_fetch_origin(ref):
-    subprocess.run(["git", "fetch", "origin", ref], stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
-
-
-def _gh_default_branch(owner_repo):
-    if not owner_repo:
-        return ""
-    rc, out, _ = _run_capture(
-        [
-            "gh",
-            "repo",
-            "view",
-            "--repo",
-            owner_repo,
-            "--json",
-            "defaultBranchRef",
-            "--jq",
-            ".defaultBranchRef.name",
-        ]
-    )
-    if rc != 0:
-        return ""
-    return (out or "").strip()
-
-
-def _git_numstat(base_ref, base_oid):
-    # Prefer commit-oid diff when available; it's unambiguous for stacked PRs.
-    candidates = []
-    if base_oid:
-        candidates.append(f"{base_oid}...HEAD")
-    if base_ref:
-        candidates.append(f"origin/{base_ref}...HEAD")
-        candidates.append(f"{base_ref}...HEAD")
-
-    for lhs in candidates:
-        rc, out, _ = _run_capture(["git", "diff", "--numstat", lhs])
-        if rc == 0:
-            return out
-    return ""
-
-
-def _parse_numstat(numstat_text):
-    rows = []
-    for line in (numstat_text or "").splitlines():
-        parts = line.split("\t")
-        if len(parts) < 3:
-            continue
-        add_s, del_s, path = parts[0].strip(), parts[1].strip(), parts[2].strip()
-        if not path:
-            continue
-        rows.append((add_s, del_s, path))
-    return rows
-
-
-def main(argv):
-    class _ArgParser(argparse.ArgumentParser):
-        def error(self, message):
-            raise ValueError(message)
-
-    parser = _ArgParser(add_help=False)
-    parser.add_argument("--pr", type=int, required=True)
-    parser.add_argument("--round", type=int, default=1)
-    try:
-        args = parser.parse_args(argv)
-    except ValueError:
-        _json_out({"error": "INVALID_ARGS"})
-        return 2
-
-    pr_number = int(args.pr)
-    round_num = int(args.round)
-
-    def _json_err(error_code, extra=None):
-        obj = {"error": error_code, "prNumber": pr_number, "round": round_num}
-        if isinstance(extra, dict) and extra:
-            obj.update(extra)
-        _json_out(obj)
-
-    # Preconditions: be in a git repo and gh is authenticated.
-    rc, out, _ = _run_capture(["git", "rev-parse", "--is-inside-work-tree"])
-    if rc != 0 or out.strip() != "true":
-        _json_err("NOT_A_GIT_REPO")
-        return 1
-
-    host = _detect_git_remote_host() or "github.com"
-    rc, gh_out, gh_err = _run_capture(["gh", "auth", "status", "--hostname", host])
-    if rc == 127:
-        _json_err(
-            "GH_CLI_NOT_FOUND",
-            {
-                "detail": "gh not found in PATH",
-                "suggestion": "Install GitHub CLI: https://cli.github.com/",
-            },
-        )
-        return 1
-    if rc != 0:
-        # If host detection is wrong, a global check might still succeed.
-        rc2, gh_out2, gh_err2 = _run_capture(["gh", "auth", "status"])
-        if rc2 == 0:
-            pass
-        else:
-            detail = (gh_err or gh_out or "").strip()
-            if len(detail) > 4000:
-                detail = detail[-4000:]
-            _json_err(
-                "GH_NOT_AUTHENTICATED",
-                {
-                    "host": host,
-                    "detail": detail,
-                    "suggestion": f"Run: gh auth login --hostname {host}",
-                },
-            )
-            return 1
-
-    rc, owner_repo, _ = _run_capture(["gh", "repo", "view", "--json", "nameWithOwner", "--jq", ".nameWithOwner"])
-    owner_repo = owner_repo.strip() if rc == 0 else ""
-    if not owner_repo:
-        _json_err("REPO_NOT_FOUND")
-        return 1
-
-    fields = "number,url,title,body,isDraft,labels,baseRefName,headRefName,baseRefOid,headRefOid,comments"
-    rc, pr_json, _ = _run_capture(["gh", "pr", "view", str(pr_number), "--repo", owner_repo, "--json", fields])
-    if rc != 0:
-        _json_err("PR_NOT_FOUND_OR_NO_ACCESS")
-        return 1
-    try:
-        pr = json.loads(pr_json)
-    except Exception:
-        _json_err("PR_NOT_FOUND_OR_NO_ACCESS")
-        return 1
-
-    head_oid = (pr.get("headRefOid") or "").strip()
-    base_oid = (pr.get("baseRefOid") or "").strip()
-    base_ref = (pr.get("baseRefName") or "").strip()
-
-    if not head_oid:
-        _json_err("PR_HEAD_OID_NOT_FOUND")
-        return 1
-
-    head_short = head_oid[:7]
-    if not base_ref:
-        base_ref = _gh_default_branch(owner_repo)
-    if not base_ref and not base_oid:
-        _json_err("PR_BASE_REF_NOT_FOUND")
-        return 1
-    head_ref = (pr.get("headRefName") or "").strip()
-    url = (pr.get("url") or "").strip()
-
-    run_id = f"{pr_number}-{round_num}-{head_short}"
-
-    if base_ref:
-        _git_fetch_origin(base_ref)
-    file_rows = _parse_numstat(_git_numstat(base_ref, base_oid))
-
-    labels = []
-    for l in (pr.get("labels") or []):
-        if isinstance(l, dict) and l.get("name"):
-            labels.append(str(l.get("name")))
-
-    comments = pr.get("comments") or []
-    recent = comments[-10:] if isinstance(comments, list) else []
-    marker_count = 0
-    for c in recent:
-        if not isinstance(c, dict):
-            continue
-        body = c.get("body") or ""
-        if isinstance(body, str) and MARKER_SUBSTR in body:
-            marker_count += 1
-
-    CACHE_DIR.mkdir(parents=True, exist_ok=True)
-    context_file = f"pr-context-pr{pr_number}-r{round_num}-{run_id}.md"
-    context_path = CACHE_DIR / context_file
-
-    with open(context_path, "w", encoding="utf-8", newline="\n") as fp:
-        fp.write("# PR Context\n\n")
-        fp.write(f"- Repo: {owner_repo}\n")
-        fp.write(f"- PR: #{pr_number} {url}\n")
-        fp.write(f"- Round: {round_num}\n")
-        fp.write(f"- RunId: {run_id}\n")
-        fp.write(f"- Base: {base_ref}\n")
-        fp.write(f"- Head: {head_ref}\n")
-        fp.write(f"- HeadShort: {head_short}\n")
-        fp.write(f"- HeadOid: {head_oid}\n")
-        fp.write(f"- Draft: {pr.get('isDraft')}\n")
-        fp.write(f"- Labels: {', '.join(labels) if labels else '(none)'}\n")
-        fp.write(f"- ExistingLoopMarkers: {marker_count}\n\n")
-
-        fp.write("## Title\n\n")
-        fp.write(_clip(pr.get("title") or "", 200) + "\n\n")
-
-        fp.write("## Body (excerpt)\n\n")
-        fp.write(_clip(pr.get("body") or "", 2000) or "(empty)")
-        fp.write("\n\n")
-
-        fp.write(f"## Changed Files ({len(file_rows)})\n\n")
-        if file_rows:
-            for add_s, del_s, path in file_rows:
-                fp.write(f"- +{add_s} -{del_s} {path}\n")
-        else:
-            fp.write("(none)\n")
-        fp.write("\n")
-
-        fp.write("## Recent Comments (excerpt)\n\n")
-        if recent:
-            for c in recent:
-                if not isinstance(c, dict):
-                    continue
-                author = None
-                if isinstance(c.get("author"), dict):
-                    author = (c.get("author") or {}).get("login")
-                fp.write(f"- {author or 'unknown'}: {_clip(c.get('body') or '', 300)}\n")
-        else:
-            fp.write("(none)\n")
-
-    _json_out(
-        {
-            "agent": "pr-context",
-            "prNumber": pr_number,
-            "round": round_num,
-            "runId": run_id,
-            "repo": {"nameWithOwner": owner_repo},
-            "headOid": head_oid,
-            "headShort": head_short,
-            "existingMarkerCount": marker_count,
-            # Handoff should be repo-relative path so downstream agents can read it directly.
-            "contextFile": _repo_relpath(REPO_ROOT, context_path),
-        }
-    )
-    return 0
-
-
-if __name__ == "__main__":
-    try:
-        raise SystemExit(main(sys.argv[1:]))
-    except SystemExit:
-        raise
-    except Exception:
-        _json_out({"error": "PR_CONTEXT_SCRIPT_FAILED"})
-        raise SystemExit(1)