ouroboros-ai 0.26.2.dev34__tar.gz → 0.26.2.dev36__tar.gz

{ouroboros_ai-0.26.2.dev34 → ouroboros_ai-0.26.2.dev36}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ouroboros-ai
-Version: 0.26.2.dev34
+Version: 0.26.2.dev36
 Summary: Specification-first workflow engine for AI coding agents. Works with Claude Code and Codex CLI.
 Author-email: Q00 <jqyu.lee@gmail.com>
 License-File: LICENSE
@@ -78,26 +78,12 @@ Description-Content-Type: text/markdown
 
 ---
 
-> **Security Notice (2026-03-24): LiteLLM PyPI Supply Chain Attack**
+> **New: PM Mode** — `ooo pm` generates a PRD through a guided PM-focused interview.
+> Ouroboros now supports product management workflows: stakeholder alignment, user story mapping, and structured PRD generation — all driven by the same Socratic engine.
 >
-> LiteLLM versions **1.82.7** and **1.82.8** on PyPI contained a malicious `.pth` file that auto-executes on every Python startup, exfiltrating environment variables, API keys, and credentials to an external server.
->
-> **What we did:** Ouroboros now pins LiteLLM to **`<=1.82.6`** (safe versions only) as an optional dependency. LiteLLM is no longer a core dependency — it is only installed when you explicitly opt in via `pip install ouroboros-ai[litellm]`.
->
-> **If you previously installed litellm 1.82.7 or 1.82.8**, the malicious `.pth` file may still be on your system even after upgrading. Clean up manually:
->
-> ```bash
-> # 1. Find and delete the malicious file
-> find / -name "litellm_init.pth" 2>/dev/null
->
-> # 2. Clean package manager caches
-> uv cache clean litellm    # if using uv
-> pip cache purge            # if using pip
->
-> # 3. Rotate ALL credentials that were on your system during the exposure window
 > ```
->
-> References: [LiteLLM #24512](https://github.com/BerriAI/litellm/issues/24512) · [Ouroboros #195](https://github.com/Q00/ouroboros/issues/195) · [DSPy mitigation](https://github.com/stanfordnlp/dspy/commit/66e33399bf29)
+> > ooo pm "I want to build a notification system"
+> ```
 
 ---
 
@@ -148,6 +134,7 @@ Then run `ooo setup` inside a Claude Code session.
 ```bash
 pip install ouroboros-ai                # base
 pip install ouroboros-ai[claude]        # + Claude Code deps
+pip install ouroboros-ai[litellm]       # + LiteLLM multi-provider
 pip install ouroboros-ai[all]           # everything
 ouroboros setup                         # configure runtime
 ```
@@ -259,6 +246,10 @@ Inside AI coding agent sessions, use `ooo <cmd>` skills. From the terminal, use
 | `ooo ralph` | *(via MCP)* | Persistent loop until verified |
 | `ooo tutorial` | *(interactive)* | Interactive hands-on learning |
 | `ooo help` | `ouroboros --help` | Full reference |
+| `ooo pm` | *(via MCP)* | PM-focused interview + PRD generation |
+| `ooo qa` | *(via skill)* | General-purpose QA verdict for any artifact |
+| `ooo update` | `ouroboros update` | Check for updates + upgrade to latest |
+| `ooo brownfield` | *(via skill)* | Scan and manage brownfield repo defaults |
 
 > Not all skills have direct CLI equivalents. Some (`evaluate`, `evolve`, `unstuck`, `ralph`) are available through agent skills or MCP tools only.
 

{ouroboros_ai-0.26.2.dev34 → ouroboros_ai-0.26.2.dev36}/README.ko.md

@@ -37,30 +37,12 @@
 
 ---
 
-> **보안 공지 (2026-03-24): LiteLLM PyPI 공급망 공격**
+> **New: PM 모드** — `ooo pm`으로 소크라테스식 PM 인터뷰를 통해 PRD를 자동 생성할 수 있습니다.
+> 이해관계자 정렬, 유저 스토리 매핑, 구조화된 PRD 생성까지 — 같은 소크라테스 엔진이 제품 관리 워크플로우도 지원합니다.
 >
-> LiteLLM 1.82.7 / 1.82.8 버전에 악성 `.pth` 파일이 포함되어 배포되었습니다. 이 파일은 `import` 없이도 **Python 시작 시 자동 실행**되며, 환경변수, API 키, SSH 키, 클라우드 크레덴셜을 외부 서버로 전송합니다.
->
-> **영향 범위:** 2026년 3월 23일 이후에 아래 중 하나라도 해당되면 감염 가능성이 있습니다:
-> - Ouroboros를 신규 설치 (`pip install ouroboros-ai`, `uv tool install ouroboros-ai`)
-> - 기존 설치를 업그레이드 (`pip install --upgrade ouroboros-ai`)
->
-> 두 경우 모두 `litellm>=1.80.0` 의존성이 악성 버전 1.82.7/1.82.8로 resolve됩니다. 그 이전에 설치하고 재설치/업그레이드하지 않았다면 안전합니다.
->
-> **Ouroboros는 litellm 의존성을 완전 제거했습니다** (v0.25.2 stable / v0.26.0b7 beta). 단, **Ouroboros를 업데이트하면 의존성 목록에서 litellm이 빠질 뿐, 이미 설치된 litellm 패키지와 악성 `.pth` 파일은 그대로 남아있습니다.** 수동 정리가 필요합니다:
->
-> ```bash
-> # 1. 악성 파일 찾기 — 결과가 나오면 즉시 삭제
-> find / -name "litellm_init.pth" 2>/dev/null
->
-> # 2. 패키지 매니저 캐시 정리
-> uv cache clean litellm    # uv 사용자
-> pip cache purge            # pip 사용자
->
-> # 3. 노출 기간 동안 시스템에 있던 모든 크레덴셜 로테이션
 > ```
->
-> 참고: [LiteLLM #24512](https://github.com/BerriAI/litellm/issues/24512) · [Ouroboros #195](https://github.com/Q00/ouroboros/issues/195) · [분석 보고서](https://futuresearch.ai/blog/litellm-pypi-supply-chain-attack/)
+> > ooo pm "알림 시스템을 만들고 싶어"
+> ```
 
 ---
 
@@ -136,6 +118,7 @@ Claude Code 세션 안에서 `ooo setup` 실행.
 ```bash
 pip install ouroboros-ai                # 기본
 pip install ouroboros-ai[claude]        # + Claude Code 의존성
+pip install ouroboros-ai[litellm]       # + LiteLLM 멀티 프로바이더
 pip install ouroboros-ai[all]           # 전부
 ouroboros setup                         # 런타임 설정
 ```
@@ -283,6 +266,11 @@ Gen 3: {Task, Priority, Status, DueDate}     → similarity 1.00 → CONVERGED
 | `ooo ralph` | 검증될 때까지 계속 도는 루프 |
 | `ooo tutorial` | 대화형 실습 |
 | `ooo help` | 전체 참조 |
+| `ooo pm` | PM 인터뷰 + PRD 생성 |
+| `ooo qa` | 범용 QA 판정 |
+| `ooo cancel` | 멈춘 실행 취소 |
+| `ooo update` | 최신 버전 확인 + 업그레이드 |
+| `ooo brownfield` | 기존 저장소 스캔 + 기본값 관리 |
 
 ---
 

{ouroboros_ai-0.26.2.dev34 → ouroboros_ai-0.26.2.dev36}/README.md

@@ -38,26 +38,12 @@
 
 ---
 
-> **Security Notice (2026-03-24): LiteLLM PyPI Supply Chain Attack**
+> **New: PM Mode** — `ooo pm` generates a PRD through a guided PM-focused interview.
+> Ouroboros now supports product management workflows: stakeholder alignment, user story mapping, and structured PRD generation — all driven by the same Socratic engine.
 >
-> LiteLLM versions **1.82.7** and **1.82.8** on PyPI contained a malicious `.pth` file that auto-executes on every Python startup, exfiltrating environment variables, API keys, and credentials to an external server.
->
-> **What we did:** Ouroboros now pins LiteLLM to **`<=1.82.6`** (safe versions only) as an optional dependency. LiteLLM is no longer a core dependency — it is only installed when you explicitly opt in via `pip install ouroboros-ai[litellm]`.
->
-> **If you previously installed litellm 1.82.7 or 1.82.8**, the malicious `.pth` file may still be on your system even after upgrading. Clean up manually:
->
-> ```bash
-> # 1. Find and delete the malicious file
-> find / -name "litellm_init.pth" 2>/dev/null
->
-> # 2. Clean package manager caches
-> uv cache clean litellm    # if using uv
-> pip cache purge            # if using pip
->
-> # 3. Rotate ALL credentials that were on your system during the exposure window
 > ```
->
-> References: [LiteLLM #24512](https://github.com/BerriAI/litellm/issues/24512) · [Ouroboros #195](https://github.com/Q00/ouroboros/issues/195) · [DSPy mitigation](https://github.com/stanfordnlp/dspy/commit/66e33399bf29)
+> > ooo pm "I want to build a notification system"
+> ```
 
 ---
 
@@ -108,6 +94,7 @@ Then run `ooo setup` inside a Claude Code session.
 ```bash
 pip install ouroboros-ai                # base
 pip install ouroboros-ai[claude]        # + Claude Code deps
+pip install ouroboros-ai[litellm]       # + LiteLLM multi-provider
 pip install ouroboros-ai[all]           # everything
 ouroboros setup                         # configure runtime
 ```
@@ -219,6 +206,10 @@ Inside AI coding agent sessions, use `ooo <cmd>` skills. From the terminal, use
 | `ooo ralph` | *(via MCP)* | Persistent loop until verified |
 | `ooo tutorial` | *(interactive)* | Interactive hands-on learning |
 | `ooo help` | `ouroboros --help` | Full reference |
+| `ooo pm` | *(via MCP)* | PM-focused interview + PRD generation |
+| `ooo qa` | *(via skill)* | General-purpose QA verdict for any artifact |
+| `ooo update` | `ouroboros update` | Check for updates + upgrade to latest |
+| `ooo brownfield` | *(via skill)* | Scan and manage brownfield repo defaults |
 
 > Not all skills have direct CLI equivalents. Some (`evaluate`, `evolve`, `unstuck`, `ralph`) are available through agent skills or MCP tools only.
 

{ouroboros_ai-0.26.2.dev34 → ouroboros_ai-0.26.2.dev36}/scripts/keyword-detector.py

@@ -19,7 +19,11 @@ import sys
 
 # Skills that work without MCP setup (bypass the setup gate)
 # qa has a built-in fallback that adopts the qa-judge agent directly
-SETUP_BYPASS_SKILLS = ["/ouroboros:setup", "/ouroboros:help", "/ouroboros:qa"]
+SETUP_BYPASS_SKILLS = [
+    "/ouroboros:setup",
+    "/ouroboros:help",
+    "/ouroboros:qa",
+]
 
 # Keyword → skill mapping
 # "ooo <cmd>" prefix always works; natural language keywords also supported
@@ -37,9 +41,22 @@ KEYWORD_MAP = [
     {"patterns": ["ooo welcome"], "skill": "/ouroboros:welcome"},
     {"patterns": ["ooo setup"], "skill": "/ouroboros:setup"},
     {"patterns": ["ooo help"], "skill": "/ouroboros:help"},
+    {"patterns": ["ooo pm", "ooo prd"], "skill": "/ouroboros:pm"},
     {"patterns": ["ooo qa", "qa check", "quality check"], "skill": "/ouroboros:qa"},
+    {"patterns": ["ooo cancel", "ooo abort"], "skill": "/ouroboros:cancel"},
     {"patterns": ["ooo update", "ooo upgrade"], "skill": "/ouroboros:update"},
+    {"patterns": ["ooo brownfield"], "skill": "/ouroboros:brownfield"},
     # Natural language triggers
+    # PM triggers must precede generic interview to avoid "pm interview" being shadowed
+    {
+        "patterns": [
+            "write prd",
+            "pm interview",
+            "product requirements",
+            "create prd",
+        ],
+        "skill": "/ouroboros:pm",
+    },
     {
         "patterns": [
             "interview me",
@@ -97,6 +114,22 @@ KEYWORD_MAP = [
         "patterns": ["update ouroboros", "upgrade ouroboros"],
         "skill": "/ouroboros:update",
     },
+    {
+        "patterns": [
+            "cancel execution",
+            "stop job",
+            "kill stuck",
+            "abort execution",
+        ],
+        "skill": "/ouroboros:cancel",
+    },
+    {
+        "patterns": [
+            "brownfield defaults",
+            "brownfield scan",
+        ],
+        "skill": "/ouroboros:brownfield",
+    },
 ]
 
 

{ouroboros_ai-0.26.2.dev34 → ouroboros_ai-0.26.2.dev36}/skills/help/SKILL.md

@@ -41,6 +41,11 @@ Ouroboros is a **requirement crystallization engine** for AI workflows. It trans
 | `ooo welcome` | First-touch welcome guide | Plugin |
 | `ooo tutorial` | Interactive hands-on learning | Plugin |
 | `ooo help` | This reference guide | Plugin |
+| `ooo pm` | PM-focused interview + PRD generation | MCP |
+| `ooo qa` | General-purpose QA verdict for any artifact | Plugin |
+| `ooo cancel` | Cancel stuck or orphaned executions | CLI |
+| `ooo update` | Check for updates + upgrade to latest | Plugin |
+| `ooo brownfield` | Scan and manage brownfield repo defaults | MCP |
 
 ### Evolutionary Loop
 
@@ -63,6 +68,16 @@ Ouroboros is a **requirement crystallization engine** for AI workflows. It trans
 | "think sideways", "i'm stuck", "break through", "lateral thinking" | `ooo unstuck` |
 | "am I drifting?", "drift check", "session status" | `ooo status` |
 
+### Utility Triggers
+
+| Phrase | Triggers |
+|--------|----------|
+| "write prd", "pm interview", "product requirements", "create prd" | `ooo pm` |
+| "qa check", "quality check" | `ooo qa` |
+| "cancel execution", "stop job", "kill stuck", "abort execution" | `ooo cancel` |
+| "update ouroboros", "upgrade ouroboros" | `ooo update` |
+| "brownfield defaults", "brownfield scan" | `ooo brownfield` |
+
 ### Loop Triggers
 
 | Phrase | Triggers |
@@ -86,6 +101,11 @@ Ouroboros is a **requirement crystallization engine** for AI workflows. It trans
 | `/ouroboros:setup` | Installation wizard | Plugin |
 | `/ouroboros:tutorial` | Interactive hands-on learning | Plugin |
 | `/ouroboros:help` | This guide | Plugin |
+| `/ouroboros:pm` | PM-focused interview + PRD generation | MCP |
+| `/ouroboros:qa` | General-purpose QA verdict for any artifact | Plugin |
+| `/ouroboros:cancel` | Cancel stuck or orphaned executions | CLI |
+| `/ouroboros:update` | Check for updates + upgrade to latest | Plugin |
+| `/ouroboros:brownfield` | Scan and manage brownfield repo defaults | MCP |
 
 ### Loop Skills
 

{ouroboros_ai-0.26.2.dev34 → ouroboros_ai-0.26.2.dev36}/src/ouroboros/_version.py

@@ -18,7 +18,7 @@ version_tuple: tuple[int | str, ...]
 commit_id: str | None
 __commit_id__: str | None
 
-__version__ = version = '0.26.2.dev34'
-__version_tuple__ = version_tuple = (0, 26, 2, 'dev34')
+__version__ = version = '0.26.2.dev36'
+__version_tuple__ = version_tuple = (0, 26, 2, 'dev36')
 
 __commit_id__ = commit_id = None

{ouroboros_ai-0.26.2.dev34 → ouroboros_ai-0.26.2.dev36}/src/ouroboros/core/worktree.py

@@ -298,8 +298,11 @@ def _pid_is_alive(pid: int) -> bool:
         return False
     try:
         os.kill(pid, 0)
-    except OSError:
+    except ProcessLookupError:
         return False
+    except PermissionError:
+        # Process exists but we lack permission to signal it — still alive.
+        return True
     return True
 
 

{ouroboros_ai-0.26.2.dev34 → ouroboros_ai-0.26.2.dev36}/src/ouroboros/mcp/tools/definitions.py

@@ -18,8 +18,6 @@ Handler modules:
 
 from __future__ import annotations
 
-from ouroboros.core.worktree import is_git_repo, maybe_restore_task_workspace  # noqa: F401
-from ouroboros.mcp.tools import evolution_handlers as _evolution_handlers
 from ouroboros.mcp.tools.authoring_handlers import (
     GenerateSeedHandler,
     InterviewHandler,
@@ -31,12 +29,10 @@ from ouroboros.mcp.tools.evaluation_handlers import (
 )
 from ouroboros.mcp.tools.evolution_handlers import (
     EvolveRewindHandler,
+    EvolveStepHandler,
     LineageStatusHandler,
     StartEvolveStepHandler,
 )
-from ouroboros.mcp.tools.evolution_handlers import (
-    EvolveStepHandler as _BaseEvolveStepHandler,
-)
 from ouroboros.mcp.tools.execution_handlers import (
     ExecuteSeedHandler,
     StartExecuteSeedHandler,
@@ -60,21 +56,6 @@ from ouroboros.mcp.tools.query_handlers import (
 # ---------------------------------------------------------------------------
 
 
-class EvolveStepHandler(_BaseEvolveStepHandler):
-    """Compatibility wrapper that respects patched helpers re-exported here."""
-
-    async def handle(self, arguments):
-        original_is_git_repo = _evolution_handlers.is_git_repo
-        original_restore_workspace = _evolution_handlers.maybe_restore_task_workspace
-        _evolution_handlers.is_git_repo = is_git_repo
-        _evolution_handlers.maybe_restore_task_workspace = maybe_restore_task_workspace
-        try:
-            return await super().handle(arguments)
-        finally:
-            _evolution_handlers.is_git_repo = original_is_git_repo
-            _evolution_handlers.maybe_restore_task_workspace = original_restore_workspace
-
-
 def execute_seed_handler(
     *,
     runtime_backend: str | None = None,

{ouroboros_ai-0.26.2.dev34 → ouroboros_ai-0.26.2.dev36}/src/ouroboros/mcp/tools/execution_handlers.py

@@ -417,6 +417,7 @@ class ExecuteSeedHandler:
                     _seed_content: str,
                     _resume_existing: bool,
                     _skip_qa: bool,
+                    _workspace: TaskWorkspace | None = workspace,
                     _session_repo: SessionRepository = session_repo,
                     _event_store: EventStore = event_store,
                     _owns_event_store: bool = owns_event_store,
@@ -481,6 +482,8 @@ class ExecuteSeedHandler:
                         except Exception:
                             log.exception("mcp.tool.execute_seed.mark_failed_error")
                     finally:
+                        if _workspace is not None:
+                            release_lock(_workspace.lock_path)
                         if _owns_event_store:
                             try:
                                 await _event_store.close()

{ouroboros_ai-0.26.2.dev34 → ouroboros_ai-0.26.2.dev36}/tests/unit/mcp/tools/test_qa_integration.py

@@ -337,7 +337,7 @@ class TestEvolveStepHandlerQA:
 
         with (
             patch(
-                "ouroboros.mcp.tools.definitions.maybe_restore_task_workspace",
+                "ouroboros.mcp.tools.evolution_handlers.maybe_restore_task_workspace",
                 return_value=None,
             ),
             patch(
@@ -378,7 +378,7 @@ class TestEvolveStepHandlerQA:
 
         with (
             patch(
-                "ouroboros.mcp.tools.definitions.maybe_restore_task_workspace",
+                "ouroboros.mcp.tools.evolution_handlers.maybe_restore_task_workspace",
                 return_value=None,
             ),
             patch(
@@ -434,7 +434,7 @@ class TestEvolveStepHandlerQA:
 
         with (
             patch(
-                "ouroboros.mcp.tools.definitions.maybe_restore_task_workspace",
+                "ouroboros.mcp.tools.evolution_handlers.maybe_restore_task_workspace",
                 return_value=None,
             ),
             patch(
@@ -467,7 +467,7 @@ class TestEvolveStepHandlerQA:
 
         with (
             patch(
-                "ouroboros.mcp.tools.definitions.maybe_restore_task_workspace",
+                "ouroboros.mcp.tools.evolution_handlers.maybe_restore_task_workspace",
                 return_value=None,
             ),
             patch(

{ouroboros_ai-0.26.2.dev34 → ouroboros_ai-0.26.2.dev36}/tests/unit/test_evolve_step.py

@@ -721,7 +721,7 @@ class TestEvolveStepHandler:
         import yaml
 
         with patch(
-            "ouroboros.mcp.tools.definitions.maybe_restore_task_workspace",
+            "ouroboros.mcp.tools.evolution_handlers.maybe_restore_task_workspace",
             return_value=None,
         ):
             result = await handler.handle(
@@ -821,9 +821,9 @@ class TestEvolveStepHandler:
         handler = EvolveStepHandler(evolutionary_loop=loop)
 
         with (
-            patch("ouroboros.mcp.tools.definitions.is_git_repo", return_value=True),
+            patch("ouroboros.mcp.tools.evolution_handlers.is_git_repo", return_value=True),
             patch(
-                "ouroboros.mcp.tools.definitions.maybe_restore_task_workspace",
+                "ouroboros.mcp.tools.evolution_handlers.maybe_restore_task_workspace",
                 side_effect=WorktreeError("Invalid durable task identifier for git worktree"),
             ),
         ):