17ba5d723b
Python rewrite of the agent harness on top of deepagents 0.6.1 + langchain 1.x, replacing the abandoned TS attempt in packages/. 388 unit/integration tests pass. Steps ----- 0. Scaffolding — uv workspace, ruff/mypy/pre-commit/alembic, src/tests/docs trees with docs/schemas/ seeded from my-deepagent-seed/. 1. Core — config (pydantic-settings with MYDEEPAGENT_ env prefix and TOML source), enums (Backend, Capability, RiskLevel, ApprovalDecisionAction, ApprovalState, RunState, RunPhaseState, SessionState, ErrorClass), errors (MyDeepAgentError + BudgetExhaustedError with PEP-3134 cause + context suppression), hash (canonical JSON + sha256). 2. Persona/Workflow/Binding — pydantic v2 schemas with tuple-based deep immutability (post-construction hash drift prevented), YAML loaders, deterministic auto-select (preferred_backends → version → name → hash), override resolution with ineligibility diagnostics, PersonaConsentStore with fcntl.flock + tmp+fsync+rename atomic write. 3. Artifact schema registry — Draft202012Validator, multi-root resolution, structured ValidationFinding output. 4. Persistence — 18 SQLAlchemy 2.0 async ORM models with FK CASCADE/RESTRICT, WAL + busy_timeout + foreign_keys PRAGMA, alembic baseline + ux_active_run_repo_base partial unique index, LangGraph SqliteSaver as context manager only (lifecycle safety). 5. DeepAgent session — build_agent wires Persona → create_deep_agent with LocalShellBackend / FilesystemBackend / StateBackend / CompositeBackend, ChatOpenAI(base_url=openrouter) for openrouter: model strings, and 4 middleware classes (cost / audit-tool / safety-shell / fallback-model). Critical workarounds -------------------- - deepagents 0.6.1 rejects FilesystemPermission together with backends that implement SandboxBackendProtocol (LocalShellBackend). SafetyShellMiddleware enforces destructive-command and secret-path policy at the tool layer instead, and build_agent strips the permissions kwarg when the persona's deepagents_backend is local_shell. - FilesystemOperation in deepagents is Literal['read', 'write'] only; _map_operations collapses our richer schema (read/write/edit/ls) safely. Real OpenRouter smoke --------------------- test_openrouter_deepagents_local_shell_smoke calls DeepSeek via deepagents + LocalShellBackend + SafetyShellMiddleware end-to-end. PASS, ~$0.000001 cost, input=9 / output=1 tokens with content "OK". Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
109 lines
3.6 KiB
YAML
109 lines
3.6 KiB
YAML
name: bug-fix-with-reproduction
|
|
version: 1
|
|
description: "버그 재현 → 진단 → 수정 → 검증. 각 단계 artifact 생성."
|
|
roles:
|
|
- id: reproducer
|
|
required_capabilities:
|
|
- evidence_check
|
|
preferred_backends:
|
|
- openrouter
|
|
fallback_personas:
|
|
- "openrouter-claude-debugger@1"
|
|
- "openrouter-deepseek-log-analyzer@1"
|
|
- id: debugger
|
|
required_capabilities:
|
|
- code_edit
|
|
- evidence_check
|
|
- command_execute
|
|
preferred_backends:
|
|
- openrouter
|
|
fallback_personas:
|
|
- "openrouter-claude-debugger@1"
|
|
- id: fixer
|
|
required_capabilities:
|
|
- code_edit
|
|
- test_first_development
|
|
preferred_backends:
|
|
- openrouter
|
|
fallback_personas:
|
|
- "openrouter-claude-code-editor@1"
|
|
- id: verifier
|
|
required_capabilities:
|
|
- evidence_check
|
|
- objective_eval
|
|
preferred_backends:
|
|
- openrouter
|
|
fallback_personas:
|
|
- "openrouter-deepseek-verifier@1"
|
|
phases:
|
|
- key: reproduce
|
|
title: "버그 재현 및 재현 조건 문서화"
|
|
risk: low
|
|
role: reproducer
|
|
expected_artifact:
|
|
path: artifacts/reproduction.json
|
|
schema: dev/spec@1
|
|
gates:
|
|
- reproduce_approved
|
|
timeout_seconds: 300
|
|
instructions: |
|
|
보고된 버그를 재현하고 재현 조건을 문서화합니다.
|
|
로그 파일이 있으면 read_file로 읽고 패턴을 분석합니다.
|
|
glob/grep으로 관련 코드를 검색합니다.
|
|
재현 조건·환경·입력값·실제 출력·기대 출력을 dev/spec@1 형식으로
|
|
artifacts/reproduction.json에 write_file로 저장합니다.
|
|
max_budget_usd: 0.20
|
|
- key: diagnose
|
|
title: "근본 원인 진단"
|
|
risk: low
|
|
role: debugger
|
|
expected_artifact:
|
|
path: artifacts/diagnosis.json
|
|
schema: dev/spec@1
|
|
gates:
|
|
- diagnose_approved
|
|
timeout_seconds: 360
|
|
instructions: |
|
|
artifacts/reproduction.json을 read_file로 읽고 근본 원인을 진단합니다.
|
|
가설을 세우고 read_file/grep으로 코드에서 검증합니다.
|
|
가장 단순한 가설부터 검증합니다.
|
|
root cause, 영향 범위, 수정 제안을 dev/spec@1 형식으로
|
|
artifacts/diagnosis.json에 write_file로 저장합니다.
|
|
max_budget_usd: 0.50
|
|
- key: fix
|
|
title: "버그 수정"
|
|
risk: medium
|
|
role: fixer
|
|
expected_artifact:
|
|
path: artifacts/fix.json
|
|
schema: dev/spec@1
|
|
gates:
|
|
- fix_approved
|
|
timeout_seconds: 600
|
|
instructions: |
|
|
artifacts/diagnosis.json을 read_file로 읽고 근본 원인을 수정합니다.
|
|
수정 전 테스트 케이스를 먼저 작성합니다 (test_first_development).
|
|
edit_file로 최소한의 변경만 적용합니다.
|
|
수정 내용, 변경된 파일 목록, 테스트 명령어를 dev/spec@1 형식으로
|
|
artifacts/fix.json에 write_file로 저장합니다.
|
|
max_budget_usd: 1.00
|
|
- key: verify
|
|
title: "수정 결과 검증"
|
|
risk: low
|
|
role: verifier
|
|
expected_artifact:
|
|
path: artifacts/verification.json
|
|
schema: dev/review-finding-batch@1
|
|
gates:
|
|
- verify_approved
|
|
timeout_seconds: 300
|
|
instructions: |
|
|
artifacts/fix.json을 read_file로 읽고 수정된 코드를 직접 확인합니다.
|
|
재현 조건이 해소됐는지, 회귀 위험은 없는지 검증합니다.
|
|
검증 결과를 dev/review-finding-batch@1 형식으로
|
|
artifacts/verification.json에 write_file로 저장합니다.
|
|
verifierStatus: confirmed = 수정 확인됨, rejected = 수정 불충분.
|
|
max_budget_usd: 0.20
|
|
default_gates: []
|
|
max_total_budget_usd: 3.0
|