bingo-light 2.1.2 → 2.2.0

package/README.en.md

@@ -10,7 +10,7 @@
   <a href="LICENSE"><img src="https://img.shields.io/badge/License-MIT-blue.svg" alt="License: MIT"></a>
   <a href="https://github.com/DanOps-1/bingo-light/releases"><img src="https://img.shields.io/github/v/release/DanOps-1/bingo-light?label=Release&color=orange" alt="Release"></a>
   <br>
-  <a href="#for-ai-agents"><img src="https://img.shields.io/badge/MCP_Server-35_tools-blueviolet.svg" alt="MCP: 35 tools"></a>
+  <a href="#for-ai-agents"><img src="https://img.shields.io/badge/MCP_Server-49_tools-blueviolet.svg" alt="MCP: 49 tools"></a>
   <a href="https://www.python.org/"><img src="https://img.shields.io/badge/Python-3.8+-3776ab.svg" alt="Python 3.8+"></a>
   <img src="https://img.shields.io/badge/Dependencies-Zero-brightgreen.svg" alt="Zero deps">
   <a href="https://github.com/DanOps-1/bingo-light/stargazers"><img src="https://img.shields.io/github/stars/DanOps-1/bingo-light?style=social" alt="Stars"></a>
@@ -25,7 +25,7 @@ GitHub's "Sync fork" button breaks the moment you have customizations. `git reba
 
 Your patches live as a clean, named stack on top of upstream. Syncing is `bingo-light sync`. Conflicts get remembered so you never solve the same one twice. And if something goes sideways, `bingo-light undo` puts everything back in one second.
 
-Every command speaks JSON. The built-in MCP server gives AI agents 35 tools to manage your fork autonomously -- from init through conflict resolution. No human in the loop required.
+Every command speaks JSON. The built-in MCP server gives AI agents 49 tools to manage your fork autonomously -- from init through conflict resolution. No human in the loop required.
 
 ---
 
@@ -69,6 +69,16 @@ That's it. Three commands and your fork stays in sync forever.
 
 > The AI calls `conflict-analyze --json`, reads the structured ours/theirs data, writes the merged file, and the rebase continues. No human needed.
 
+During a rebase, `conflict-analyze` returns a full situational briefing:
+- **`patch_intent`** — patch name, subject, full commit message, original SHA, original diff, metadata, stack position.
+- **`verify`** — configured `test.command` + per-file syntax/parse hints by extension (`.py/.json/.yml/.yaml/.toml/.sh`).
+- **`upstream_context`** — upstream commits touching conflicts, with author, subject, and extracted PR numbers.
+- **`patch_dependencies`** — later patches in your stack that modify overlapping files (cascade risk).
+- **`decision_memory`** — prior resolutions for this patch from `.bingo/decisions/`, ranked by relevance.
+- Each `conflicts[]` entry carries **`semantic_class`**: `whitespace` / `import_reorder` / `signature_change` / `logic`.
+
+`conflict-resolve --verify` (CLI) or `verify: true` (MCP) runs `test.command` after the final `git rebase --continue`; the result is attached as `verify_result`. `conflict-resolve` also auto-records the decision (file, semantic class, strategy) to decision memory.
+
 ---
 
 ## Key Features
@@ -93,7 +103,7 @@ That's it. Three commands and your fork stays in sync forever.
 
 ### For AI Agents
 
-- :electric_plug: **MCP server (35 tools)** -- full fork management from init through conflict resolution.
+- :electric_plug: **MCP server (49 tools)** -- full fork management from init through conflict resolution.
 - :bar_chart: **`--json` on everything** -- every command returns structured JSON. Parse, don't scrape.
 - :mute: **`--yes` flag** -- fully non-interactive. No TTY required. No prompts. Ever.
 - :gear: **Auto-detect non-TTY** -- pipes and subprocesses trigger non-interactive mode automatically.
@@ -200,7 +210,7 @@ make install && bingo-light setup
 
 ## For AI Agents
 
-bingo-light was designed from day one for AI agents. Every command speaks JSON. The MCP server exposes 35 tools covering the full lifecycle from `init` to `conflict-resolve`. Non-interactive mode is the default when stdin is not a TTY.
+bingo-light was designed from day one for AI agents. Every command speaks JSON. The MCP server exposes 49 tools covering the full lifecycle from `init` to `conflict-resolve`. Non-interactive mode is the default when stdin is not a TTY.
 
 ### MCP setup -- Claude Code
 
@@ -234,7 +244,7 @@ Add to `~/Library/Application Support/Claude/claude_desktop_config.json`:
 
 **Any MCP client** (VS Code Copilot, Cursor, custom agents): connect via stdio to `python3 mcp-server.py`.
 
-### 35 MCP Tools
+### 49 MCP Tools
 
 | Tool | Purpose |
 |------|---------|
@@ -424,7 +434,7 @@ StGit (649 stars) manages patch stacks but has no AI integration, no MCP server,
 | Handles customizations | **Yes** | **No** | Manual | Manual | Manual |
 | Conflict memory (rerere) | **Auto** | No | Manual | No | No |
 | Conflict prediction | **Yes** | No | No | No | No |
-| AI / MCP integration | **35 tools** | No | No | No | No |
+| AI / MCP integration | **49 tools** | No | No | No | No |
 | JSON output | **All commands** | No | No | No | No |
 | Non-interactive mode | **Native** | No | Partial | Partial | Partial |
 | Undo sync | **One command** | No | git reflog | Manual | Manual |
@@ -490,7 +500,7 @@ Yes. bingo-light uses standard git operations (fetch, rebase, push). It works wi
 ```
 bingo-light          CLI tool (Python 3, zero deps)
 bingo_core/          Core library package (all business logic)
-mcp-server.py        MCP server (zero-dep Python 3, 35 tools, JSON-RPC 2.0)
+mcp-server.py        MCP server (zero-dep Python 3, 49 tools, JSON-RPC 2.0)
 contrib/agent.py     Advisor agent (monitors drift, auto-syncs when safe)
 contrib/tui.py       Terminal dashboard (curses TUI, real-time monitoring)
 install.sh           Installer (--yes for CI, --help for options)

package/README.md

@@ -9,7 +9,7 @@
   <a href="https://github.com/DanOps-1/bingo-light/actions"><img src="https://github.com/DanOps-1/bingo-light/actions/workflows/ci.yml/badge.svg" alt="CI"></a>
   <a href="LICENSE"><img src="https://img.shields.io/badge/License-MIT-blue.svg" alt="License: MIT"></a>
   <a href="https://github.com/DanOps-1/bingo-light/releases"><img src="https://img.shields.io/github/v/release/DanOps-1/bingo-light?label=Release&color=orange" alt="Release"></a>
-  <a href="#mcp-服务器"><img src="https://img.shields.io/badge/MCP_Server-35_tools-blueviolet.svg" alt="MCP: 35 tools"></a>
+  <a href="#mcp-服务器"><img src="https://img.shields.io/badge/MCP_Server-49_tools-blueviolet.svg" alt="MCP: 49 tools"></a>
   <a href="https://www.python.org/"><img src="https://img.shields.io/badge/Python-3.8+-3776ab.svg" alt="Python 3.8+"></a>
   <img src="https://img.shields.io/badge/Dependencies-Zero-brightgreen.svg" alt="Zero deps">
   <a href="https://github.com/DanOps-1/bingo-light/stargazers"><img src="https://img.shields.io/github/stars/DanOps-1/bingo-light?style=social" alt="Stars"></a>
@@ -168,10 +168,36 @@ cd bingo-light && make install && bingo-light setup
       "theirs": "... 你的补丁版本 ...",
       "hint": "上游重构了调度器核心；补丁需要适配新结构。"
     }
-  ]
+  ],
+  "patch_intent": {
+    "name": "custom-scheduler",
+    "subject": "...",
+    "message": "完整 commit 消息",
+    "original_sha": "a1b2c3d...",
+    "original_diff": "diff --git ...",
+    "meta": {"reason": "...", "tags": [], "status": "permanent"},
+    "stack_position": {"index": 3, "total": 7}
+  },
+  "verify": {
+    "test_command": "make test",
+    "file_hints": [
+      {"file": "kernel/sched/core.c", "command": "bash -n ...", "kind": "syntax"}
+    ]
+  }
 }
 ```
 
+`conflict-analyze` 在 rebase 中返回完整态势简报：
+
+- **`patch_intent`**：补丁意图（原始 commit、diff、metadata、栈位置）
+- **`verify`**：配置的 `test.command` + 按扩展名的逐文件校验命令
+- **`upstream_context`**：触发冲突的上游 commits（作者、主题、自动抽取的 PR 号）
+- **`patch_dependencies`**：栈中后续补丁是否触及相同文件（cascade 风险）
+- **`decision_memory`**：该补丁的历史解决方案（`.bingo/decisions/`）
+- 每条 `conflicts[]` 都带 **`semantic_class`**：`whitespace` / `import_reorder` / `signature_change` / `logic`
+
+`conflict-resolve --verify` 在最终 `git rebase --continue` 完成后自动跑 `test.command`，结果挂在 `verify_result` 字段；`conflict-resolve` 还会自动把本次决策（file, semantic class, strategy）写入 decision memory。
+
 </details>
 
 <details>
@@ -197,7 +223,7 @@ Claude Code:
 ```console
 $ bingo-light setup
 
-  ◆  bingo-light setup  v2.1.1
+  ◆  bingo-light setup  v2.x.x
   │
   ◆  MCP Server
   │  Connect bingo-light tools to your AI coding assistants
@@ -434,7 +460,7 @@ bingo-light help                             打印帮助
 
 | 集成方式 | 适用场景 | 示例 |
 |---------|---------|------|
-| **MCP** (35 tools) | Claude Code / Cursor / Windsurf 等 | `bingo-light setup` 自动配 |
+| **MCP** (49 tools) | Claude Code / Cursor / Windsurf 等 | `bingo-light setup` 自动配 |
 | **CLI `--json`** | 任何能跑 shell 的 AI | `bingo-light sync --json --yes` |
 | **Skill** | Claude Code / Continue / Gemini 等 | `/bingo` 教 AI 用法 |
 

package/bingo-light

@@ -234,6 +234,7 @@ def _format_conflict_analyze(result: dict) -> str:
     conflicts = result.get("conflicts", [])
     if not conflicts:
         return f"{_c(GREEN, 'OK')} No conflicts detected."
+
     lines = [f"{_c(YELLOW, '!')} {len(conflicts)} conflicting file(s):"]
     for c in conflicts:
         f = c.get("file", "?")
@@ -241,6 +242,85 @@ def _format_conflict_analyze(result: dict) -> str:
         lines.append(f"  {_c(RED, f)}")
         if hint:
             lines.append(f"    hint: {hint}")
+
+    intent = result.get("patch_intent")
+    if intent and intent.get("name"):
+        lines.append("")
+        lines.append(
+            f"{_c(BOLD, 'Patch intent:')} "
+            f"{intent['name']} \u2014 {intent.get('subject', '')}"
+        )
+        meta = intent.get("meta") or {}
+        if meta.get("reason"):
+            lines.append(f"  reason: {meta['reason']}")
+        if meta.get("tags"):
+            lines.append(f"  tags: {', '.join(meta['tags'])}")
+        sp = intent.get("stack_position")
+        if sp:
+            lines.append(f"  position: {sp['index']}/{sp['total']}")
+
+    verify = result.get("verify")
+    if verify:
+        lines.append("")
+        lines.append(f"{_c(BOLD, 'Verify suggestions:')}")
+        tc = verify.get("test_command")
+        if tc:
+            lines.append(f"  test.command: {tc}")
+        for h in verify.get("file_hints", []):
+            lines.append(f"  [{h['kind']}] {h['command']}")
+
+    deps = result.get("patch_dependencies")
+    if deps and deps.get("dependents"):
+        lines.append("")
+        dlist = deps["dependents"]
+        lines.append(
+            f"{_c(BOLD, 'Cascade risk:')} "
+            f"{len(dlist)} later patch(es) touch overlapping files"
+        )
+        for d in dlist[:5]:
+            files = ", ".join(d.get("overlapping_files", []))
+            lines.append(
+                f"  #{d['position']} {d['name']}: {d.get('subject', '')} "
+                f"\u2014 {_c(YELLOW, files)}"
+            )
+        if len(dlist) > 5:
+            lines.append(f"  {_c(DIM, f'... {len(dlist) - 5} more')}")
+
+    dm = result.get("decision_memory")
+    if dm and dm.get("entries"):
+        lines.append("")
+        lines.append(f"{_c(BOLD, 'Decision memory:')} prior resolutions for this patch")
+        for e in dm["entries"]:
+            lines.append(f"  {_c(CYAN, e['file'])} ({e['semantic_class']}):")
+            for prev in e.get("previous_decisions", [])[:3]:
+                strat = prev.get("resolution_strategy", "?")
+                subj = prev.get("upstream_subject") or "(no subject)"
+                ts = prev.get("timestamp", "")
+                rel = prev.get("relevance", "")
+                lines.append(
+                    f"    [{strat}] {subj} {_c(DIM, f'({rel}, {ts})')}"
+                )
+
+    uc = result.get("upstream_context")
+    if uc:
+        lines.append("")
+        total = uc.get("total_commits", 0)
+        touching = uc.get("commits_touching_conflicts", [])
+        lines.append(
+            f"{_c(BOLD, 'Upstream context:')} "
+            f"{uc.get('range', '?')} "
+            f"({total} commits, {len(touching)} touching conflicts)"
+        )
+        for commit in touching[:5]:
+            short = commit.get("short_sha", "?")
+            subj = commit.get("subject", "")
+            author = commit.get("author", "")
+            pr = commit.get("pr")
+            pr_tag = f" #{pr}" if pr else ""
+            lines.append(f"  {_c(DIM, short)} {subj}{pr_tag} \u2014 {author}")
+        if len(touching) > 5:
+            lines.append(f"  {_c(DIM, f'... {len(touching) - 5} more')}")
+
     return "\n".join(lines)
 
 
@@ -261,10 +341,18 @@ def _format_conflict_resolve(result: dict) -> str:
         )
 
     if result.get("sync_complete"):
-        return (
-            f"{_c(GREEN, 'OK')} Resolved {resolved} "
-            f"\u2014 sync complete!"
-        )
+        out = f"{_c(GREEN, 'OK')} Resolved {resolved} \u2014 sync complete!"
+        vr = result.get("verify_result")
+        if vr:
+            if vr.get("skipped"):
+                out += f"\n  verify: skipped ({vr.get('reason', '')})"
+            elif vr.get("test") == "pass":
+                out += f"\n  {_c(GREEN, 'verify: PASS')} ({vr.get('command', '')})"
+            elif vr.get("test") == "fail":
+                out += f"\n  {_c(RED, 'verify: FAIL')} ({vr.get('command', '')})"
+            else:
+                out += f"\n  verify: {vr.get('test', '?')}"
+        return out
 
     if result.get("conflict"):
         patch = result.get("current_patch", "")
@@ -284,10 +372,18 @@ def _format_conflict_resolve(result: dict) -> str:
         return "\n".join(lines)
 
     if result.get("rebase_continued"):
-        return (
-            f"{_c(GREEN, 'OK')} Resolved {resolved} "
-            f"\u2014 rebase continued."
-        )
+        out = f"{_c(GREEN, 'OK')} Resolved {resolved} \u2014 rebase continued."
+        vr = result.get("verify_result")
+        if vr:
+            if vr.get("skipped"):
+                out += f"\n  verify: skipped ({vr.get('reason', '')})"
+            elif vr.get("test") == "pass":
+                out += f"\n  {_c(GREEN, 'verify: PASS')} ({vr.get('command', '')})"
+            elif vr.get("test") == "fail":
+                out += f"\n  {_c(RED, 'verify: FAIL')} ({vr.get('command', '')})"
+            else:
+                out += f"\n  verify: {vr.get('test', '?')}"
+        return out
 
     return f"{_c(GREEN, 'OK')} Resolved {resolved}."
 
@@ -500,6 +596,118 @@ def _format_patch_meta(result: dict) -> str:
     return f"  {k}: {v}" if k else f"{_c(GREEN, 'OK')} Done."
 
 
+def _format_patch_lock(result: dict) -> str:
+    """Format patch lock/unlock result."""
+    if result.get("ok") is False:
+        return f"{_c(RED, 'x')} {result.get('error', 'Failed.')}"
+    patch = result.get("patch", "")
+    owner = result.get("owner", "")
+    if result.get("was_locked") is False:
+        return f"{_c(GREEN, 'OK')} {patch} was not locked."
+    if "locked_at" in result:
+        reason = result.get("reason", "")
+        msg = f"{_c(GREEN, 'OK')} Locked {_c(BOLD, patch)} by {owner}"
+        if reason:
+            msg += f" ({reason})"
+        return msg
+    prev = result.get("previous_owner", "")
+    if prev and prev != owner:
+        return f"{_c(GREEN, 'OK')} Unlocked {_c(BOLD, patch)} (was locked by {prev})"
+    return f"{_c(GREEN, 'OK')} Unlocked {_c(BOLD, patch)}"
+
+
+def _format_patch_check(result: dict) -> str:
+    """Format patch check (obsolescence detection) result."""
+    if result.get("ok") is False:
+        return f"{_c(RED, 'x')} {result.get('error', 'Failed.')}"
+    patches = result.get("patches", [])
+    if not patches:
+        return "No patches to check."
+    lines = []
+    for p in patches:
+        status = p.get("status", "unknown")
+        name = p.get("name", "?")
+        reason = p.get("reason", "")
+        if status == "obsolete":
+            lines.append(f"  {_c(YELLOW, 'OBSOLETE')} {_c(BOLD, name)} — {reason}")
+        elif status == "active":
+            lines.append(f"  {_c(GREEN, 'ACTIVE')}   {_c(BOLD, name)} — {reason}")
+        else:
+            lines.append(f"  {_c(DIM, 'UNKNOWN')}  {name} — {reason}")
+    header = f"Checked {len(patches)} patch(es):"
+    return header + "\n" + "\n".join(lines)
+
+
+def _format_patch_upstream(result: dict) -> str:
+    """Format patch upstream (PR-ready export) result."""
+    if result.get("ok") is False:
+        return f"{_c(RED, 'x')} {result.get('error', 'Failed.')}"
+    name = result.get("patch", "")
+    desc = result.get("description", "")
+    files = result.get("files", [])
+    stats = result.get("stats", "")
+    diff = result.get("diff", "")
+    lines = [
+        f"{_c(BOLD, 'PR-ready export:')} {name}",
+        f"  Description: {desc}",
+        f"  Files: {', '.join(files)}",
+    ]
+    if stats:
+        lines.append(f"  Stats: {stats}")
+    lines.append("")
+    lines.append(diff)
+    return "\n".join(lines)
+
+
+def _format_patch_expire(result: dict) -> str:
+    """Format patch expire result."""
+    if result.get("ok") is False:
+        return f"{_c(RED, 'x')} {result.get('error', 'Failed.')}"
+    expired = result.get("expired", [])
+    expiring = result.get("expiring_soon", [])
+    if not expired and not expiring:
+        return f"{_c(GREEN, 'OK')} No expired or expiring patches."
+    lines = []
+    if expired:
+        lines.append(f"{_c(RED, 'Expired')} ({len(expired)}):")
+        for e in expired:
+            lines.append(f"  {_c(RED, 'x')} {_c(BOLD, e['name'])} — expired {e['expires']} ({abs(e['days_left'])}d ago)")
+    if expiring:
+        lines.append(f"{_c(YELLOW, 'Expiring soon')} ({len(expiring)}):")
+        for e in expiring:
+            lines.append(f"  {_c(YELLOW, '!')} {_c(BOLD, e['name'])} — expires {e['expires']} ({e['days_left']}d left)")
+    return "\n".join(lines)
+
+
+def _format_patch_stats(result: dict) -> str:
+    """Format patch stats result."""
+    if result.get("ok") is False:
+        return f"{_c(RED, 'x')} {result.get('error', 'Failed.')}"
+    patches = result.get("patches", [])
+    if not patches:
+        return "No patches."
+    lines = [f"{'#':>3}  {'Name':<20} {'Age':>6} {'Size':>10} {'Syncs':>6} {'Status':<12} {'Owner':<15}"]
+    lines.append("─" * 80)
+    for i, p in enumerate(patches, 1):
+        age = f"{p['age_days']}d" if p.get("age_days", -1) >= 0 else "?"
+        size = f"+{p.get('insertions', 0)}/-{p.get('deletions', 0)}"
+        syncs = str(p.get("syncs_survived", 0))
+        status = p.get("status", "")
+        owner = p.get("owner", "") or p.get("locked_by", "") or ""
+        lines.append(f"{i:>3}  {p['name']:<20} {age:>6} {size:>10} {syncs:>6} {status:<12} {owner:<15}")
+    return "\n".join(lines)
+
+
+def _format_report(result: dict) -> str:
+    """Format fork health report."""
+    if result.get("ok") is False:
+        return f"{_c(RED, 'x')} {result.get('error', 'Failed.')}"
+    report = result.get("report", "")
+    if report:
+        return report
+    return f"{_c(GREEN, 'OK')} Report generated."
+
+
 def _format_undo(result: dict) -> str:
     """Format undo result."""
     if result.get("ok") is False:
@@ -740,7 +948,11 @@ def build_parser() -> argparse.ArgumentParser:
     sub.add_parser("diff", aliases=["d"], add_help=False)
 
     # doctor
-    sub.add_parser("doctor", add_help=False)
+    p_doctor = sub.add_parser("doctor", add_help=False)
+    p_doctor.add_argument("--report", action="store_true")
+
+    # report
+    sub.add_parser("report", add_help=False)
 
     # log
     sub.add_parser("log", add_help=False)
@@ -755,6 +967,8 @@ def build_parser() -> argparse.ArgumentParser:
     cr = sub.add_parser("conflict-resolve", add_help=False)
     cr.add_argument("resolve_file", nargs="?", default="")
     cr.add_argument("--content-stdin", action="store_true")
+    cr.add_argument("--verify", action="store_true",
+                    help="Run test.command after the final rebase continues")
 
     # session
     p_session = sub.add_parser("session", add_help=False)
@@ -811,6 +1025,23 @@ def build_parser() -> argparse.ArgumentParser:
     pp_meta.add_argument("meta_key", nargs="?", default="")
     pp_meta.add_argument("meta_value", nargs="?", default="")
 
+    pp_lock = patch_sub.add_parser("lock", add_help=False)
+    pp_lock.add_argument("target")
+    pp_lock.add_argument("--reason", default="")
+
+    pp_unlock = patch_sub.add_parser("unlock", add_help=False)
+    pp_unlock.add_argument("target")
+    pp_unlock.add_argument("--force", action="store_true")
+
+    pp_check = patch_sub.add_parser("check", add_help=False)
+    pp_check.add_argument("target", nargs="?", default="")
+
+    pp_upstream = patch_sub.add_parser("upstream", add_help=False)
+    pp_upstream.add_argument("target")
+
+    patch_sub.add_parser("expire", add_help=False)
+    patch_sub.add_parser("stats", add_help=False)
+
     # workspace
     p_ws = sub.add_parser("workspace", aliases=["ws"], add_help=False)
     ws_sub = p_ws.add_subparsers(dest="ws_command")
@@ -853,6 +1084,26 @@ def build_parser() -> argparse.ArgumentParser:
     dp_drop.add_argument("dep_package")
     dp_drop.add_argument("dep_patch_name", nargs="?", default="")
 
+    # dep override subcommands
+    dp_override = dep_sub.add_parser("override", add_help=False)
+    override_sub = dp_override.add_subparsers(dest="override_command")
+    override_sub.add_parser("list", add_help=False)
+    override_sub.add_parser("check", add_help=False)
+    ov_add = override_sub.add_parser("add", add_help=False)
+    ov_add.add_argument("override_package")
+    ov_add.add_argument("override_version")
+    ov_add.add_argument("--reason", dest="override_reason", default="")
+    ov_drop = override_sub.add_parser("drop", add_help=False)
+    ov_drop.add_argument("override_package")
+
+    # dep fork subcommands
+    dp_fork = dep_sub.add_parser("fork", add_help=False)
+    fork_sub = dp_fork.add_subparsers(dest="fork_command")
+    fork_sub.add_parser("list", add_help=False)
+    fork_sub.add_parser("check", add_help=False)
+    fk_sync = fork_sub.add_parser("sync", add_help=False)
+    fk_sync.add_argument("fork_package")
+
     # help
     sub.add_parser("help", add_help=False)
 
@@ -884,6 +1135,31 @@ def dispatch(args: argparse.Namespace, json_mode: bool) -> Optional[dict]:
             return dm.sync()
         if dcmd == "drop":
             return dm.drop(args.dep_package, getattr(args, "dep_patch_name", ""))
+        if dcmd == "override":
+            ocmd = getattr(args, "override_command", "")
+            if ocmd == "list":
+                return dm.override_list()
+            if ocmd == "check":
+                return dm.override_check()
+            if ocmd == "add":
+                return dm.override_add(
+                    args.override_package, args.override_version,
+                    getattr(args, "override_reason", ""),
+                )
+            if ocmd == "drop":
+                return dm.override_drop(args.override_package)
+            return {"ok": False, "error": f"Unknown override subcommand: {ocmd}"}
+        if dcmd == "fork":
+            from bingo_core.dep_fork import ForkTracker
+            ft = ForkTracker()
+            fcmd = getattr(args, "fork_command", "")
+            if fcmd == "list":
+                return ft.fork_list()
+            if fcmd == "check":
+                return ft.fork_check()
+            if fcmd == "sync":
+                return ft.fork_sync(args.fork_package)
+            return {"ok": False, "error": f"Unknown fork subcommand: {fcmd}"}
         return {"ok": False, "error": f"Unknown dep subcommand: {dcmd}"}
 
     # setup doesn't need a Repo (works outside git repos)
@@ -919,7 +1195,10 @@ def dispatch(args: argparse.Namespace, json_mode: bool) -> Optional[dict]:
         return repo.diff()
 
     if cmd == "doctor":
-        return repo.doctor()
+        return repo.doctor(report=getattr(args, "report", False))
+
+    if cmd == "report":
+        return repo.report()
 
     if cmd == "log":
         return repo.history()
@@ -935,7 +1214,9 @@ def dispatch(args: argparse.Namespace, json_mode: bool) -> Optional[dict]:
         if args.content_stdin:
             import sys as _sys
             content = _sys.stdin.read()
-        return repo.conflict_resolve(args.resolve_file, content)
+        return repo.conflict_resolve(
+            args.resolve_file, content, verify=args.verify
+        )
 
     if cmd == "session":
         update = (args.session_action == "update")
@@ -1018,6 +1299,24 @@ def _dispatch_patch(args: argparse.Namespace, repo: Repo) -> dict:
             value=meta_value,
         )
 
+    if pcmd == "lock":
+        return repo.patch_lock(args.target, reason=getattr(args, "reason", ""))
+
+    if pcmd == "unlock":
+        return repo.patch_unlock(args.target, force=getattr(args, "force", False))
+
+    if pcmd == "check":
+        return repo.patch_check(getattr(args, "target", ""))
+
+    if pcmd == "upstream":
+        return repo.patch_upstream(args.target)
+
+    if pcmd == "expire":
+        return repo.patch_expire()
+
+    if pcmd == "stats":
+        return repo.patch_stats()
+
     raise BingoError(f"Unknown patch subcommand: {pcmd}")
 
 
@@ -1075,6 +1374,7 @@ _FORMATTERS: Dict[str, Any] = {
     "smart-sync": _format_smart_sync,
     "setup": _format_setup,
     "dep": _format_dep,
+    "report": _format_report,
 }
 
 
@@ -1099,6 +1399,16 @@ def _get_formatter(args: argparse.Namespace):
             return _format_patch_import
         if pcmd == "meta":
             return _format_patch_meta
+        if pcmd in ("lock", "unlock"):
+            return _format_patch_lock
+        if pcmd == "check":
+            return _format_patch_check
+        if pcmd == "upstream":
+            return _format_patch_upstream
+        if pcmd == "expire":
+            return _format_patch_expire
+        if pcmd == "stats":
+            return _format_patch_stats
         return _format_generic
 
     # workspace subcommands

package/bingo_core/__init__.py

@@ -14,7 +14,7 @@ import re
 
 # --- Constants ---
 
-VERSION = "2.1.2"
+VERSION = "2.2.0"
 PATCH_PREFIX = "[bl]"
 CONFIG_FILE = ".bingolight"
 BINGO_DIR = ".bingo"
@@ -39,9 +39,12 @@ from bingo_core.exceptions import (  # noqa: E402
     DirtyTreeError,
 )
 from bingo_core.models import PatchInfo, ConflictInfo  # noqa: E402
+from bingo_core.semantic import classify_conflict  # noqa: E402
+from bingo_core.decisions import DecisionMemory, detect_resolution_strategy  # noqa: E402
 from bingo_core.git import Git  # noqa: E402
 from bingo_core.config import Config  # noqa: E402
 from bingo_core.state import State  # noqa: E402
+from bingo_core.team import TeamState  # noqa: E402
 from bingo_core.repo import Repo  # noqa: E402
 
 __all__ = [
@@ -69,9 +72,14 @@ __all__ = [
     # Data classes
     "PatchInfo",
     "ConflictInfo",
+    # Functions
+    "classify_conflict",
+    "detect_resolution_strategy",
     # Classes
+    "DecisionMemory",
     "Git",
     "Config",
     "State",
+    "TeamState",
     "Repo",
 ]