본문으로 건너뛰기
cccskills
무료GitHub에 공개

geode-code-conventions

Apply GEODE's measured architecture, module and symbol naming, Python type/class, persisted schema, test, site, and versioning conventions. Use when adding or moving modules, choosing dataclass/Pydantic/Protocol/ABC forms, naming classes or files, designing stored data, reviewing convention compliance, or deciding compatibility and release version treatment.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

설치 방법 보기

포함된 파일(2)

  • SKILL.md6.7 KB
  • agents/openai.yaml267 B

SKILL.md(원문)

설치하기 전에 에이전트에게 전달되는 지시 내용을 확인할 수 있습니다.

GEODE Code Conventions

Use this skill to make a convention decision against the current repository, not against a generic Python style guide.

Canonical reference

docs/architecture/naming-conventions.md is the detailed convention SOT. Do not copy its tables into this skill.

Progressive disclosure

Always read the SOT introduction, evidence snapshot, and review checklist. Then read only the sections that match the change:

DecisionRead
requirements, plans, acceptance, backend claims, or scaffold removalRequirements, plans, and evidence
ownership, layers, or abstraction level§1 Architecture and §6 Imports
package, file, function, class, tool, event, or React naming§2 Packages, files, and symbols
public hooks, middleware, or runtime eventsHook contracts: canonical names, decision authority, callers, and regression checks
Python data form, typing, construction, or lifecycle§3 Python typing and class design
persisted data, schema identity, or writer/reader compatibility§4 Data, schemas, and persistence
errors, logging, redaction, or trust boundaries§5 Errors, logging, and trust boundaries
test placement or verification design§7 Tests and verification
Next.js or public-site code§8 Site conventions
package version, compatibility, or promotion§9 Versioning and compatibility

Follow every applicable row for a cross-cutting change. Read the whole SOT only when auditing or changing the convention system itself.

Executable configuration remains authoritative:

  • pyproject.toml for Ruff, mypy, pytest, import-linter, dependency, security, and coverage gates;
  • AGENTS.md and docs/workflow.md for repository workflow;
  • site/DESIGN.md, site/tsconfig.json, and site/eslint.config.mjs for the public site;
  • the subsystem contract nearest the changed code for local invariants.

Workflow

  1. Locate ownership. Identify the top-level root and the package that owns the decision. Confirm allowed dependency direction before naming anything.
  2. Inspect siblings. Search the nearest canonical modules, their tests, and their public imports. Classify an apparent pattern as required, preferred, compatibility-only, or historical.
  3. Choose the smallest form. Prefer a function, focused module, dataclass, or concrete class before adding a Protocol, registry, factory, or package.
  4. Separate identities. Keep module names, Python types, persisted schema IDs, protocol versions, algorithm versions, and package versions as distinct decisions.
  5. Apply symmetrically. Update imports, exports, tests, registries, schema validators, docs, and generated mirrors that share the contract. Do not add an alias unless a released consumer needs the documented grace period.
  6. Audit the diff. Search for the old name, parent-relative imports, generic catch-all names, unbounded Any, hidden mutation, missing teardown, stale schema IDs, and version stamp drift.
  7. Verify proportionally. Run the targeted behavior test and the exact static/build gates for every touched language and layer. Report what did not run.

Decision rules

  • Package context removes redundant filename prefixes; persisted artifact identity does not change merely because a module moved.
  • Protocol is the normal structural port. Use ABC only for a shared template algorithm with real subclass hooks.
  • Use frozen/slotted dataclasses for immutable internal values, Pydantic at validated external/config/LLM boundaries, TypedDict for deliberately dict-shaped bridges, and StrEnum for reusable closed persisted vocabularies.
  • Keep Any at the untyped boundary and narrow it immediately.
  • Use absolute imports across boundaries; allow only level-one relative imports inside a tightly coupled local package.
  • Give every persisted record one authority, schema identity, validator, and version axis. Writers emit current schemas; legacy readers are explicit.
  • Treat observed zero/false/empty values separately from missing None/null.
  • Keep direct output in CLIs and scripts; library modules use structured, redacted logging.
  • Prefer async-native runtime I/O and explicit ownership of offloaded work; apply §§3.3 and 5 to cancellation, terminal persistence, IPC error propagation, auth authority, and bounded diagnostic fields.
  • Mirror production package paths in tests and prove wiring pairs together.
  • Routine post-1.0 releases are PATCH; MINOR/MAJOR require operator approval.

Specialist routing

Use this skill with, rather than instead of:

NeedSkill or SOT
implementation and GitFlow$geode-workflow
dependency cycles or layer violations$dependency-review
Python quality and lifecycle review$code-review-quality
schema/log/event/trajectory consistency$geode-workflow observability reference
evaluation artifacts$geode-eval
model-facing prompt text$prompt-writing
changelog or release version$geode-changelog

Output contract

State the ownership, selected convention, local precedent, compatibility impact, and verification evidence. If deviating, name the exact rule, explain why the local contract is stronger, and keep the exception scoped to the introducing change.

리뷰

아직 리뷰가 없습니다. 사용해 본 소감을 남겨 주세요.

댓글

로그인 없이 쓸 수 있습니다. 사용 후기, 질문, 작동하지 않은 환경 등을 편하게 남겨 주세요.

아직 댓글이 없습니다. 첫 댓글을 남겨 보세요.

보내면 바로 공개됩니다. 이메일 주소나 전화번호 같은 개인 정보는 쓰지 마세요. URL은 2개까지입니다.

0 / 1,000자

이 양식은 reCAPTCHA로 보호되며 Google 개인정보처리방침 및 서비스 약관이 적용됩니다.

같은 저장소의 스킬

Audit agent runtimes and their tests for lifecycle inference, capability overload, context pollution, correlated verification, prompt-only control, false success, dead compatibility paths, test-only seams, and complexity without contribution. Use for whole-repository agent audits, slop or dead-code triage, test contribution reviews, large refactors, tool/context simplification, and cleanup requests that need proof before deleting code or tests.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

mangowhoiscloud/geode152026년 10월 16일 업데이트

Diagnose agent operational failures through default-policy checks, multi-gap analysis, explicit service ownership, request-context lifetime, and degradation-versus-correctness checks.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

mangowhoiscloud/geode152026년 10월 16일 업데이트

Align, run, audit, and publish GEODE benchmark comparisons against Agent-World arXiv 2604.18292v1 across MCP-Mark, BFCL V4, and tau2-bench. Use for Agent-World or AgentWorld comparison requests, cross-suite agent benchmark design, Tau2 comparator/profile diagnosis, paired thin-runtime-versus-GEODE experiments, mean-accuracy@8 replication, or benchmark comparability and artifact reviews.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

mangowhoiscloud/geode152026년 10월 16일 업데이트

Verification checklist to prevent fake success. Detects test deletion/disabling, coverage regression, lint bypass, secret exposure. Triggered by "deception", "fake" ("가짜"), "fake success", "verification" ("검증"), "checklist" ("체크리스트"), "deletion detection" ("삭제 탐지"), "regression" keywords.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

mangowhoiscloud/geode152026년 10월 16일 업데이트

Implement proven backend architecture patterns including Clean Architecture, Hexagonal Architecture, and Domain-Driven Design. Use when architecting complex backend systems or refactoring existing applications for better maintainability.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

mangowhoiscloud/geode152026년 10월 16일 업데이트

Auto-search and summarize latest AI/agent papers.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

mangowhoiscloud/geode152026년 10월 16일 업데이트

mangowhoiscloud의 스킬 모두 보기

이 스킬의 문제 신고