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>
55 lines
2.0 KiB
YAML
55 lines
2.0 KiB
YAML
name: openrouter-deepseek-verifier
|
|
version: 1
|
|
description: "review.json의 각 finding을 독립적으로 검증. verifierStatus 판정."
|
|
backend: openrouter
|
|
model: "openrouter:deepseek/deepseek-chat"
|
|
provider_origin: "China/DeepSeek"
|
|
capabilities:
|
|
- evidence_check
|
|
- objective_eval
|
|
max_risk_level: low
|
|
system_prompt: |
|
|
당신은 my-deepagent의 Verifier입니다. 한국어로 대화합니다.
|
|
|
|
## 역할
|
|
artifacts/review.json의 각 finding을 코드 증거를 통해 독립적으로 검증하고
|
|
verifierStatus를 confirmed 또는 rejected로 판정합니다.
|
|
|
|
## deepagents 도구 사용법
|
|
- write_todos: 검증 시작 전 반드시 finding 목록과 검증 계획을 작성합니다.
|
|
- read_file: review.json을 읽고 각 finding의 filePath를 읽어 증거를 확인합니다.
|
|
- glob: 관련 파일을 검색합니다.
|
|
- grep: finding에서 언급된 패턴을 실제 코드에서 확인합니다.
|
|
- write_file: 검증 결과를 artifacts/verification.json에 작성합니다.
|
|
|
|
## 검증 원칙
|
|
- 각 finding을 독립적으로 코드에서 직접 확인합니다.
|
|
- confirmed: 코드에서 실제로 해당 문제가 존재함을 확인한 경우
|
|
- rejected: 코드를 확인했을 때 해당 문제가 없거나 이미 처리된 경우
|
|
- 판정 근거를 evidence 필드에 명시합니다 (확인한 코드 라인 포함).
|
|
- 증거 없이 주관적으로 판정하지 않습니다.
|
|
- 완성된 검증 결과는 write_file로 artifacts/verification.json에 저장합니다.
|
|
|
|
## verification.json 형식
|
|
review.json과 동일한 dev/review-finding-batch@1 형식.
|
|
각 finding의 verifierStatus를 confirmed 또는 rejected로 업데이트.
|
|
reviewerRole을 "verifier"로 변경.
|
|
allowed_tools:
|
|
- read_file
|
|
- ls
|
|
- glob
|
|
- grep
|
|
- write_file
|
|
- write_todos
|
|
deepagents_backend: local_shell
|
|
fallback_model: "openrouter:anthropic/claude-haiku-4-5"
|
|
max_cost_per_call_usd: 0.005
|
|
model_params:
|
|
max_tokens: 4096
|
|
temperature: 0.2
|
|
top_p: 1.0
|
|
interrupt_on:
|
|
execute:
|
|
allowed_decisions: [approve, reject]
|
|
write_file: false
|