Back to Details

academic-research-suite

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Academic Research Suite for Codex

Codex专属学术研究套件

This is a Codex adapter for the ARS suite. The vendored ARS content lives under 
ars/
; keep it as source material and route through this file first.
本套件是ARS套件的Codex适配器。ARS相关内容存放在
ars/
目录下；请将其作为源素材，先通过本文件进行路由。

Versioning

版本控制

This Codex package is version 
0.1.7
. The repo-root 
VERSION
, this 
SKILL.md
metadata version, and 
manifest.json
adapter_version
must match. Vendored ARS suite versions are tracked separately by source repository commit in 
manifest.json
.
本Codex包版本为
0.1.7
。仓库根目录的
VERSION
、本
SKILL.md
的元数据版本以及
manifest.json
中的
adapter_version
必须保持一致。ARS套件的版本通过
manifest.json
中源仓库的提交记录单独跟踪。

First Rule

首要规则

Do not load the whole suite by default. Select one workflow, read that workflow's 
WORKFLOW.md
, then load only the agent, reference, template, or shared files needed for the user's current stage.
The internal workflow entry files are named 
WORKFLOW.md
, not 
SKILL.md
, so Codex registers only this root router skill instead of exposing every vendored upstream workflow as a separate skill.
默认情况下不要加载整个套件。选择一个工作流，阅读该工作流的
WORKFLOW.md
，然后仅加载用户当前阶段所需的Agent、参考资料、模板或共享文件。
内部工作流入口文件命名为
WORKFLOW.md
，而非
SKILL.md
，因此Codex仅注册此根路由技能，不会将每个上游工作流作为单独技能暴露。

Workflow Router

工作流路由

Choose the workflow by intent:
User intentRead first
Deep research, literature review, systematic review, meta-analysis, fact-checking, research question refinement
ars/deep-research/WORKFLOW.md
Academic paper writing, paper outline, abstract, revision, citation formatting, AI disclosure, LaTeX/DOCX/PDF formatting guidance
ars/academic-paper/WORKFLOW.md
Paper review, peer review simulation, editorial decision, reviewer calibration, re-review after revision
ars/academic-paper-reviewer/WORKFLOW.md
End-to-end research-to-paper pipeline, integrity gate, staged review/revision/finalization workflow
ars/academic-pipeline/WORKFLOW.md
Experiment planning, code experiment execution plan, human study protocol, statistical interpretation, reproducibility validation
ars/experiment-agent/WORKFLOW.md
If the request spans multiple workflows, start with 
ars/academic-pipeline/WORKFLOW.md
unless the user clearly asked for a single phase.
根据用户意图选择工作流：
用户意图首先阅读
深度研究、文献综述、系统综述、元分析、事实核查、研究问题细化
ars/deep-research/WORKFLOW.md
学术论文写作、论文大纲、摘要、修改、引用格式、AI披露、LaTeX/DOCX/PDF格式指导
ars/academic-paper/WORKFLOW.md
论文评审、同行评审模拟、编辑决策、审稿人校准、修改后重审
ars/academic-paper-reviewer/WORKFLOW.md
端到端研究转论文流程、完整性检查、阶段性评审/修改/定稿工作流
ars/academic-pipeline/WORKFLOW.md
实验规划、代码实验执行计划、人体研究方案、统计解读、可重复性验证
ars/experiment-agent/WORKFLOW.md
如果用户请求涉及多个工作流，除非用户明确要求单阶段处理，否则从
ars/academic-pipeline/WORKFLOW.md
开始。

Paper Topic Scoping Override

论文主题范围覆盖规则

Apply this override before the general paper/pipeline routing rule and before the Claude-Style Alias Router below. The override applies regardless of whether the user invokes ARS via natural language or via an 
ars-*
alias.
If the user says they want to write a paper, thesis, proposal, article, journal article, or manuscript, but they only provide a broad topic, tentative title, research interest, or "題目/主題/方向" and do not provide a clear, answerable research question, route to 
ars/deep-research/WORKFLOW.md
in 
socratic
mode first. This matches the upstream ARS experience where vague paper-topic requests start with SCR/Socratic narrowing instead of immediate outline or drafting.
Treat these as Socratic triggers even when the wording contains paper-writing intent:
  • "I want to write a paper on ..."
  • "I have a paper topic/title ..."
  • "我想做一篇論文，題目是..."
  • "我有一個研究方向/主題，但還不確定問題"
  • "幫我想論文題目/收斂研究問題"
First response in this path:
  1. State that the request is being routed to 
    deep-research
    socratic
    mode because the research question is not yet precise.
  2. Ask 3-5 Socratic narrowing questions using 
    socratic_mentor_agent
    and 
    research_question_agent
    guidance.
  3. Do not produce an outline, draft, literature review, or full pipeline dashboard until the user has converged on at least one candidate RQ.
Route directly to 
ars/academic-paper/WORKFLOW.md
only when the user already has a clear RQ, approved study frame, data/results, literature matrix, draft, or explicitly asks to skip scoping and proceed to outline/drafting. Route to 
ars/academic-pipeline/WORKFLOW.md
only when the user explicitly asks for the full research-to-paper pipeline or says to continue after Socratic scoping.
在通用论文/流程路由规则和下文的Claude风格别名路由规则之前，先应用此覆盖规则。无论用户是通过自然语言还是
ars-*
别名调用ARS，均适用此规则。
如果用户表示想要撰写论文、学位论文、提案、文章、期刊文章或稿件，但仅提供宽泛主题、暂定标题、研究兴趣或“題目/主題/方向”，且未提供清晰、可解答的研究问题，则先路由至
ars/deep-research/WORKFLOW.md
的
socratic
模式。这与上游ARS的使用体验一致：模糊的论文主题请求会先通过SCR/苏格拉底式收窄，而非直接进入大纲或起草阶段。
即使表述中包含论文写作意图，以下情况也视为苏格拉底式触发：
  • "I want to write a paper on ..."
  • "I have a paper topic/title ..."
  • "我想做一篇論文，題目是..."
  • "我有一個研究方向/主題，但還不確定問題"
  • "幫我想論文題目/收斂研究問題"
此路径下的首次响应：
  1. 说明由于研究问题尚不明确，请求将被路由至
    deep-research
    的
    socratic
    模式。
  2. 依据
    socratic_mentor_agent
    和
    research_question_agent
    的指导，提出3-5个苏格拉底式收窄问题。
  3. 在用户确定至少一个候选研究问题（RQ）之前，不要生成大纲、草稿、文献综述或完整流程仪表盘。
仅当用户已有清晰的研究问题、获批的研究框架、数据/结果、文献矩阵、草稿，或明确要求跳过范围界定直接进入大纲/起草阶段时，才直接路由至
ars/academic-paper/WORKFLOW.md
。仅当用户明确要求完整的研究转论文流程，或表示在苏格拉底式范围界定后继续处理时，才路由至
ars/academic-pipeline/WORKFLOW.md
。

Claude-Style Alias Router

Claude风格别名路由

Codex does not install Claude slash commands, but this package emulates their intent. If the user's request starts with a slash alias (
/ars-plan
) or a plain alias (
ars-plan
), treat it as a mode shortcut, strip the alias token from the task text, read the matching 
ars/commands/ars-*.md
prompt recipe, then route to the workflow 
WORKFLOW.md
below.
The 
model:
field in command frontmatter is a Claude routing hint only. Codex uses the current model unless the user explicitly requests another model.
AliasRead command recipeThen route to
/ars-plan
, 
ars-plan
ars/commands/ars-plan.md
ars/academic-paper/WORKFLOW.md
in 
plan
mode
/ars-outline
, 
ars-outline
ars/commands/ars-outline.md
ars/academic-paper/WORKFLOW.md
in 
outline-only
mode
/ars-abstract
, 
ars-abstract
ars/commands/ars-abstract.md
ars/academic-paper/WORKFLOW.md
in 
abstract-only
mode
/ars-lit-review
, 
ars-lit-review
ars/commands/ars-lit-review.md
ars/academic-paper/WORKFLOW.md
in 
lit-review
mode; if the user wants source discovery and synthesis instead, route to 
ars/deep-research/WORKFLOW.md
in 
lit-review
mode
/ars-citation-check
, 
ars-citation-check
ars/commands/ars-citation-check.md
ars/academic-paper/WORKFLOW.md
in 
citation-check
mode
/ars-disclosure
, 
ars-disclosure
ars/commands/ars-disclosure.md
ars/academic-paper/WORKFLOW.md
in 
disclosure
mode
/ars-format-convert
, 
ars-format-convert
ars/commands/ars-format-convert.md
ars/academic-paper/WORKFLOW.md
in 
format-convert
mode
/ars-revision-coach
, 
ars-revision-coach
ars/commands/ars-revision-coach.md
ars/academic-paper/WORKFLOW.md
in 
revision-coach
mode
/ars-revision
, 
ars-revision
ars/commands/ars-revision.md
ars/academic-paper/WORKFLOW.md
in 
revision
mode
/ars-full
, 
ars-full
ars/commands/ars-full.md
ars/academic-pipeline/WORKFLOW.md
If the request body after the alias is a vague topic, tentative title, research direction, or "題目/主題/方向" without a clear research question, defer to the Paper Topic Scoping Override above before routing to the alias's target mode. This applies to 
ars-plan
, 
ars-outline
, 
ars-abstract
, 
ars-lit-review
, and 
ars-full
.
If the Codex client reserves slash-prefixed input before it reaches the model, tell the user to use the plain alias form, for example 
ars-plan my topic
.
Codex未安装Claude斜杠命令，但本包模拟其意图。如果用户请求以斜杠别名（
/ars-plan
）或普通别名（
ars-plan
）开头，则将其视为模式快捷方式，从任务文本中移除别名标记，阅读对应的
ars/commands/ars-*.md
提示词模板，然后路由至下方的工作流
WORKFLOW.md
。
命令前置元数据中的
model:
字段仅为Claude路由提示。Codex使用当前模型，除非用户明确请求其他模型。
别名阅读命令模板然后路由至
/ars-plan
, 
ars-plan
ars/commands/ars-plan.md
ars/academic-paper/WORKFLOW.md
的
plan
模式
/ars-outline
, 
ars-outline
ars/commands/ars-outline.md
ars/academic-paper/WORKFLOW.md
的
outline-only
模式
/ars-abstract
, 
ars-abstract
ars/commands/ars-abstract.md
ars/academic-paper/WORKFLOW.md
的
abstract-only
模式
/ars-lit-review
, 
ars-lit-review
ars/commands/ars-lit-review.md
ars/academic-paper/WORKFLOW.md
的
lit-review
模式；若用户需要来源发现与合成，则路由至
ars/deep-research/WORKFLOW.md
的
lit-review
模式
/ars-citation-check
, 
ars-citation-check
ars/commands/ars-citation-check.md
ars/academic-paper/WORKFLOW.md
的
citation-check
模式
/ars-disclosure
, 
ars-disclosure
ars/commands/ars-disclosure.md
ars/academic-paper/WORKFLOW.md
的
disclosure
模式
/ars-format-convert
, 
ars-format-convert
ars/commands/ars-format-convert.md
ars/academic-paper/WORKFLOW.md
的
format-convert
模式
/ars-revision-coach
, 
ars-revision-coach
ars/commands/ars-revision-coach.md
ars/academic-paper/WORKFLOW.md
的
revision-coach
模式
/ars-revision
, 
ars-revision
ars/commands/ars-revision.md
ars/academic-paper/WORKFLOW.md
的
revision
模式
/ars-full
, 
ars-full
ars/commands/ars-full.md
ars/academic-pipeline/WORKFLOW.md
如果别名后的请求内容是模糊主题、暂定标题、研究方向或“題目/主題/方向”，且无清晰研究问题，则先遵循上述论文主题范围覆盖规则，再路由至别名的目标模式。此规则适用于
ars-plan
、
ars-outline
、
ars-abstract
、
ars-lit-review
和
ars-full
。
如果Codex客户端在请求到达模型前保留了斜杠前缀输入，请告知用户使用普通别名形式，例如
ars-plan my topic
。

Codex Runtime Mapping

Codex运行时映射

The upstream ARS files were written for Claude Code. Apply these mappings when using them in Codex:
Upstream wordingCodex behavior
Agent Team, agent, dispatch, handoffRead the referenced 
agents/*.md
file as a role or phase prompt and perform that phase inline.
Agent tool, Task tool, subagentDo not spawn agents automatically. Only use Codex subagents when the user explicitly asks for delegation or parallel agents.
AskUserQuestionAsk concise clarification questions, or use Codex's structured user-input tool when available in the active mode.
WebSearchUse Codex web browsing for current facts, source verification, citation checks, and external evidence. Provide source links.
Bash, Write, EditTreat as capability descriptions, not required tool names. Follow Codex safety rules and the user's filesystem constraints.
Claude, Claude Code, model-specific wordingInterpret as "the current Codex agent" unless the text is part of a disclosure template or historical example.
ARS_CROSS_MODEL
, 
ARS_CROSS_MODEL_SAMPLE_INTERVAL
Treat upstream secondary-model dispatch instructions as no-op unless the user explicitly asks for cross-model review. When explicitly enabled in this Codex package, use Anthropic Claude Opus 4.7 via API (
ARS_CROSS_MODEL=claude-opus-4.7
, 
ANTHROPIC_API_KEY
); do not route this reviewer through Codex/OpenAI APIs. Skip unconfigured cross-model report sections instead of inventing results.
S2_API_KEY
, 
OPENALEX_POLITE_EMAIL
, 
CROSSREF_POLITE_EMAIL
These are optional upstream bibliographic lookup settings. Use them only when the user explicitly runs contamination-signal migration or programmatic reference verification; normal Codex routing does not require them.
fresh Claude Code session
, 
Claude Code session
Read as "a new Codex conversation". Material Passport reset semantics still apply; only the runtime changes. This rule covers 
ars/academic-pipeline/WORKFLOW.md
, 
ars/academic-pipeline/agents/pipeline_orchestrator_agent.md
, 
ars/academic-pipeline/references/passport_as_reset_boundary.md
, 
ars/experiment-agent/README.md
, 
ars/experiment-agent/README.zh-TW.md
, and 
ars/docs/PERFORMANCE.md
.
/ars-*
slash command, Claude plugin command		Treat 
ars/commands/ars-*.md
as optional prompt recipes. Codex does not register slash commands from this package.
SessionStart hook, SubagentStop hook, 
hooks/hooks.json
Treat as upstream Claude Code hook metadata only. Do not install or execute Claude hooks in Codex unless the user explicitly asks to inspect or port a hook.
上游ARS文件是为Claude Code编写的。在Codex中使用时，请应用以下映射：
上游表述Codex行为
Agent Team、agent、dispatch、handoff将引用的
agents/*.md
文件视为角色或阶段提示词，并内联执行该阶段。
Agent tool、Task tool、subagent不要自动生成Agent。仅当用户明确请求委托或并行Agent时，才使用Codex子Agent。
AskUserQuestion提出简洁的澄清问题，或在当前模式可用时使用Codex的结构化用户输入工具。
WebSearch使用Codex网页浏览功能获取当前事实、来源验证、引用检查和外部证据，并提供来源链接。
Bash、Write、Edit视为功能描述，而非必需工具名称。遵循Codex安全规则和用户的文件系统限制。
Claude、Claude Code、模型特定表述解释为“当前Codex Agent”，除非文本属于披露模板或历史示例。
ARS_CROSS_MODEL
、
ARS_CROSS_MODEL_SAMPLE_INTERVAL
将上游跨模型调度指令视为无操作，除非用户明确请求跨模型评审。当在本Codex包中明确启用时，通过API使用Anthropic Claude Opus 4.7（
ARS_CROSS_MODEL=claude-opus-4.7
、
ANTHROPIC_API_KEY
）；不要通过Codex/OpenAI API路由此评审器。跳过未配置的跨模型报告部分，而非编造结果。
S2_API_KEY
、
OPENALEX_POLITE_EMAIL
、
CROSSREF_POLITE_EMAIL
这些是可选的上游文献查找设置。仅当用户明确运行污染信号迁移或程序化参考验证时使用；常规Codex路由不需要这些设置。
fresh Claude Code session
、
Claude Code session
解读为“新的Codex对话”。材料通行证重置语义仍然适用；仅运行时改变。此规则覆盖
ars/academic-pipeline/WORKFLOW.md
、
ars/academic-pipeline/agents/pipeline_orchestrator_agent.md
、
ars/academic-pipeline/references/passport_as_reset_boundary.md
、
ars/experiment-agent/README.md
、
ars/experiment-agent/README.zh-TW.md
和
ars/docs/PERFORMANCE.md
。
/ars-*
斜杠命令、Claude插件命令		将
ars/commands/ars-*.md
视为可选提示词模板。Codex不会从此包注册斜杠命令。
SessionStart hook、SubagentStop hook、
hooks/hooks.json
仅视为上游Claude Code钩子元数据。除非用户明确要求检查或移植钩子，否则不要在Codex中安装或执行Claude钩子。

Agent Prompt Use

Agent提示词使用

When a workflow lists agents:
  1. Read the workflow 
    WORKFLOW.md
    to identify the mode and phase.
  2. Read the specific 
    agents/<name>.md
    files for the current phase.
  3. Treat each agent file as a scoped role prompt with an input/output contract.
  4. Produce the phase output in the current conversation unless the user requested files.
  5. Use 
    ars/shared/handoff_schemas.md
    when a phase hands material to another phase.
For multi-review phases, preserve independence by writing each reviewer section before synthesizing. Do not let the final synthesis erase critical findings from devil's advocate or methodology roles.
当工作流列出Agent时：
  1. 阅读工作流
    WORKFLOW.md
    以确定模式和阶段。
  2. 阅读当前阶段对应的特定
    agents/<name>.md
    文件。
  3. 将每个Agent文件视为具有输入/输出契约的限定角色提示词。
  4. 在当前对话中生成阶段输出，除非用户请求文件。
  5. 当一个阶段向另一个阶段移交材料时，使用
    ars/shared/handoff_schemas.md
    。
对于多评审阶段，在合成前保留每个评审部分的独立性。不要让最终合成掩盖来自魔鬼代言人或方法论角色的关键发现。

Canonical Agent Files

标准Agent文件

Use these exact filenames. Do not invent hyphenated alternatives or rename files from memory.
ars/deep-research/agents/
: 
bibliography_agent.md
, 
devils_advocate_agent.md
, 
editor_in_chief_agent.md
, 
ethics_review_agent.md
, 
meta_analysis_agent.md
, 
monitoring_agent.md
, 
report_compiler_agent.md
, 
research_architect_agent.md
, 
research_question_agent.md
, 
risk_of_bias_agent.md
, 
socratic_mentor_agent.md
, 
source_verification_agent.md
, 
synthesis_agent.md
.
ars/academic-paper/agents/
: 
abstract_bilingual_agent.md
, 
argument_builder_agent.md
, 
citation_compliance_agent.md
, 
draft_writer_agent.md
, 
formatter_agent.md
, 
intake_agent.md
, 
literature_strategist_agent.md
, 
peer_reviewer_agent.md
, 
revision_coach_agent.md
, 
socratic_mentor_agent.md
, 
structure_architect_agent.md
, 
visualization_agent.md
.
ars/academic-paper-reviewer/agents/
: 
devils_advocate_reviewer_agent.md
, 
domain_reviewer_agent.md
, 
editorial_synthesizer_agent.md
, 
eic_agent.md
, 
field_analyst_agent.md
, 
methodology_reviewer_agent.md
, 
perspective_reviewer_agent.md
.
ars/academic-pipeline/agents/
: 
claim_ref_alignment_audit_agent.md
, 
collaboration_depth_agent.md
, 
integrity_verification_agent.md
, 
pipeline_orchestrator_agent.md
, 
state_tracker_agent.md
.
ars/experiment-agent/agents/
: 
code_runner_agent.md
, 
study_manager_agent.md
.
使用以下确切文件名。不要自行发明连字符替代名称或凭记忆重命名文件。
ars/deep-research/agents/
： 
bibliography_agent.md
、
devils_advocate_agent.md
、 
editor_in_chief_agent.md
、
ethics_review_agent.md
、 
meta_analysis_agent.md
、
monitoring_agent.md
、 
report_compiler_agent.md
、
research_architect_agent.md
、 
research_question_agent.md
、
risk_of_bias_agent.md
、 
socratic_mentor_agent.md
、
source_verification_agent.md
、 
synthesis_agent.md
。
ars/academic-paper/agents/
： 
abstract_bilingual_agent.md
、
argument_builder_agent.md
、 
citation_compliance_agent.md
、
draft_writer_agent.md
、 
formatter_agent.md
、
intake_agent.md
、 
literature_strategist_agent.md
、
peer_reviewer_agent.md
、 
revision_coach_agent.md
、
socratic_mentor_agent.md
、 
structure_architect_agent.md
、
visualization_agent.md
。
ars/academic-paper-reviewer/agents/
： 
devils_advocate_reviewer_agent.md
、
domain_reviewer_agent.md
、 
editorial_synthesizer_agent.md
、
eic_agent.md
、 
field_analyst_agent.md
、
methodology_reviewer_agent.md
、 
perspective_reviewer_agent.md
。
ars/academic-pipeline/agents/
： 
claim_ref_alignment_audit_agent.md
、
collaboration_depth_agent.md
、 
integrity_verification_agent.md
、 
pipeline_orchestrator_agent.md
、
state_tracker_agent.md
。
ars/experiment-agent/agents/
： 
code_runner_agent.md
、
study_manager_agent.md
。

Shared Resources

共享资源

Use 
ars/shared/
for cross-workflow contracts and quality gates:
  • ars/shared/handoff_schemas.md
    defines inter-stage artifact schemas.
  • ars/shared/style_calibration_protocol.md
    defines writing voice calibration.
  • ars/shared/mode_spectrum.md
    defines fidelity, balanced, and originality modes.
  • ars/shared/agents/compliance_agent.md
    defines compliance checks.
  • ars/shared/compliance_checkpoint_protocol.md
    , 
    ars/shared/prisma_trAIce_protocol.md
    , and 
    ars/shared/raise_framework.md
    define integrity and reporting gates.
  • ars/scripts/
    contains upstream validators and reference adapters.
  • ars/examples/
    contains upstream non-PDF fixtures and templates.
  • ars/docs/design/
    contains upstream design specs referenced by ARS protocols.
  • ars/commands/
    contains upstream Claude slash-command prompt recipes.
  • ars/hooks/
    contains upstream Claude hook metadata preserved for traceability.
  • ars/tests/
    contains upstream fixture corpora used by validator tests.
When an ARS file points to 
shared/...
, resolve it as 
ars/shared/...
. When it points to another workflow, resolve it under 
ars/<workflow>/...
. When it points to root-level 
scripts/...
, 
examples/...
, or 
docs/...
, resolve it under 
ars/scripts/...
, 
ars/examples/...
, or 
ars/docs/...
.
使用
ars/shared/
获取跨工作流契约和质量检查：
  • ars/shared/handoff_schemas.md
    定义了阶段间工件模式。
  • ars/shared/style_calibration_protocol.md
    定义了写作风格校准规则。
  • ars/shared/mode_spectrum.md
    定义了保真度、平衡和原创性模式。
  • ars/shared/agents/compliance_agent.md
    定义了合规性检查。
  • ars/shared/compliance_checkpoint_protocol.md
    、
    ars/shared/prisma_trAIce_protocol.md
    和
    ars/shared/raise_framework.md
    定义了完整性和报告检查。
  • ars/scripts/
    包含上游验证器和参考适配器。
  • ars/examples/
    包含上游非PDF示例和模板。
  • ars/docs/design/
    包含ARS协议引用的上游设计规范。
  • ars/commands/
    包含上游Claude斜杠命令提示词模板。
  • ars/hooks/
    包含为可追溯性保留的上游Claude钩子元数据。
  • ars/tests/
    包含验证器测试使用的上游示例语料库。
当ARS文件指向
shared/...
时，解析为
ars/shared/...
。当指向另一个工作流时，解析为
ars/<workflow>/...
。当指向根目录的
scripts/...
、
examples/...
或
docs/...
时，解析为
ars/scripts/...
、
ars/examples/...
或
ars/docs/...
。

Inactive Upstream Scripts

非活跃上游脚本

manifest.json
lists 
inactive_upstream_scripts
that are vendored for traceability but are not Codex package validation gates. Do not wire them into Codex CI or treat them as required runtime checks unless the missing upstream Claude Code inputs, especially 
.claude/CLAUDE.md
, are deliberately supplied.
ars/scripts/run_codex_audit.sh
is vendored because upstream ARS uses it as a Codex audit wrapper, but follow its own guardrail: it must not be invoked from the same in-LLM session that produced the audited deliverable.
manifest.json
列出了
inactive_upstream_scripts
，这些脚本为了可追溯性而被纳入，但不是Codex包的验证检查。除非故意提供缺失的上游Claude Code输入（尤其是
.claude/CLAUDE.md
），否则不要将其接入Codex CI或视为必需的运行时检查。
ars/scripts/run_codex_audit.sh
被纳入是因为上游ARS将其作为Codex审计包装器，但需遵循其自身的防范规则：不得在生成审计交付物的同一场LLM会话中调用它。

Verification Discipline

验证原则

For claims, citations, references, statistics, journal policies, API behavior, and current facts, verify against primary or authoritative sources. If verification is not possible, mark the item as unverified instead of inventing support.
Never fabricate references. For citation existence checks, prefer DOI or official metadata lookup, then authoritative web search. Semantic Scholar, OpenAlex, and Crossref API instructions are in 
ars/deep-research/references/
; use them only when the task needs programmatic reference verification.
对于主张、引用、参考资料、统计数据、期刊政策、API行为和当前事实，需对照原始或权威来源进行验证。如果无法验证，应标记为未验证，而非编造支持依据。
切勿伪造参考资料。对于引用存在性检查，优先使用DOI或官方元数据查找，再进行权威网页搜索。Semantic Scholar、OpenAlex和Crossref API说明位于
ars/deep-research/references/
；仅当任务需要程序化参考验证时才使用。

Output Defaults

输出默认值

  • Default language follows the user's language.
  • For Chinese, use Traditional Chinese unless the user requests otherwise.
  • For staged workflows, show the current stage, required inputs, output artifact, and whether the next gate is optional or mandatory.
  • For paper/research outputs, keep uncertainty explicit and separate evidence, inference, and recommendation.
  • 默认语言遵循用户使用的语言。
  • 中文默认使用繁体中文，除非用户另有要求。
  • 对于阶段性工作流，显示当前阶段、所需输入、输出工件，以及下一检查点是可选还是必填。
  • 对于论文/研究输出，明确保留不确定性，区分证据、推断和建议。