athena-work

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Athena Brief Processor

Athena研究包处理器

Process

.athenabrief

research packages exported from the Athena note-taking app. These are ZIP archives containing curated notes, AI-generated summaries, a research brief, and binary assets arranged for progressive disclosure.

处理从Athena笔记应用导出的

.athenabrief

研究包。这些是ZIP压缩包，包含整理后的笔记、AI生成的摘要、研究简报以及按渐进式披露方式组织的二进制资源。

Package Structure Overview

包结构概述

package.athenabrief (ZIP)
├── instructions.md        # Static reading guide (start here)
├── brief.md               # Research brief, report, and note index
├── summaries.json         # Machine-readable note summaries
├── manifest.json          # Full machine-readable index and statistics
├── references/*.md        # Full note content with YAML frontmatter
└── assets/*               # Binary attachments (images, PDFs)

Read

references/athenabrief-spec.md

for full schema details when encountering unusual fields or edge cases.

package.athenabrief (ZIP)
├── instructions.md        # 静态阅读指南（从这里开始）
├── brief.md               # 研究简报、报告和笔记索引
├── summaries.json         # 机器可读的笔记摘要
├── manifest.json          # 完整的机器可读索引和统计信息
├── references/*.md        # 带有YAML前置元数据的完整笔记内容
└── assets/*               # 二进制附件（图片、PDF等）

遇到不常见字段或边缘情况时，请阅读

references/athenabrief-spec.md

获取完整的格式说明。

Processing Workflow

处理工作流

Step 1: Extract and Orient

步骤1：提取与初步了解

Extract the

.athenabrief

ZIP to a temporary working directory. Read

instructions.md

first to understand the package's progressive disclosure model. This file is static across all packages and confirms the reading order.

将

.athenabrief

ZIP包提取到临时工作目录。首先阅读

instructions.md

以了解包的渐进式披露模型。该文件在所有包中都是静态的，用于确认阅读顺序。

Step 2: Progressive Disclosure Reads

步骤2：渐进式阅读

Read files in this order, stopping as early as possible:

brief.md
— Read the research objective, suggested web research, research report, and note index table. This provides the high-level context and tells you what the brief is about.
summaries.json
— Read the machine-readable note summaries. Each entry has
```
generalSummary
```
,
```
contextualSummary
```
,
```
keyContributions
```
, and
```
knowledgeEntries
```
. This gives you enough detail to plan your approach.
Short-circuit check — Decide whether you need deeper detail (see decision matrix below). In ~80% of cases, brief.md + summaries.json provide sufficient context.
manifest.json
— Read only when you need statistics, relevance scores, sub-query coverage, or the full research plan. Skip if summaries.json was sufficient.
references/*.md
— Read individual notes only when you need the full text. Use the manifest or summaries to identify which notes are relevant. Each note has YAML frontmatter with metadata and the full markdown body.
assets/*
— Read binary assets only when referenced in the work or explicitly requested by the user.

按照以下顺序读取文件，尽早停止以提高效率：

brief.md
— 阅读研究目标、建议的网络研究内容、研究报告和笔记索引表。这部分提供了高层级上下文，让你了解简报的核心内容。
summaries.json
— 阅读机器可读的笔记摘要。每个条目包含
```
generalSummary
```
、
```
contextualSummary
```
、
```
keyContributions
```
和
```
knowledgeEntries
```
。这些信息足以让你规划处理方法。
短路检查 — 判断是否需要更深入的细节（见下方决策矩阵）。约80%的情况下，
```
brief.md
```
+
```
summaries.json
```
已提供足够的上下文。
manifest.json
— 仅当你需要统计数据、相关性评分、子查询覆盖范围或完整研究计划时才阅读。如果
```
summaries.json
```
已足够，则跳过此文件。
references/*.md
— 仅当你需要完整文本时才阅读单个笔记。使用manifest或摘要来确定哪些笔记是相关的。每个笔记都包含带有元数据的YAML前置内容和完整的Markdown正文。
assets/*
— 仅当工作中引用或用户明确要求时，才读取二进制资源。

Step 3: Short-Circuit Decision Matrix

步骤3：短路决策矩阵

After reading brief.md and summaries.json, use this matrix:

Signal	Action
User gave specific instructions	Follow them; read references only for cited notes
Brief has clear research objective + report covers it	Summarize findings from brief.md + summaries; skip references
User asks "what's in this brief?"	Summarize from brief.md table + summaries.json; skip references
Task requires synthesis across notes	Read relevant references (use summaries to pick which)
Task requires exact quotes or data	Read the specific reference notes cited
Task requires web research	Note the `webResearchInstructions` from brief.md; search the web
No user instructions at all	Use brief's research objective as the task; process accordingly

阅读

brief.md

和

summaries.json

后，使用以下矩阵判断：

信号	操作
用户给出了具体指令	遵循指令；仅阅读引用的笔记
简报有明确的研究目标且报告已覆盖该目标	从 `brief.md` 和摘要中总结发现；跳过参考笔记
用户询问“这个简报里有什么？”	从 `brief.md` 表格和 `summaries.json` 中总结；跳过参考笔记
任务需要跨笔记综合分析	阅读相关的参考笔记（使用摘要选择相关内容）
任务需要精确引用或数据	阅读引用的特定参考笔记
任务需要网络研究	记录 `brief.md` 中的 `webResearchInstructions` ；进行网络搜索
完全没有用户指令	将简报的研究目标作为任务；进行相应处理

Step 4: Engagement Rules

步骤4：交互规则

Ask the user for clarification when ANY of these apply:

Imprecise brief — The research objective is vague and multiple interpretations exist
Fuzzy search instructions — The web research instructions are ambiguous
Insufficient knowledge — The brief references domains you cannot reliably address
Contradictory data — Notes contain conflicting information with no clear resolution
Scope ambiguity — Unclear whether the user wants a summary, analysis, new content, or something else

When the brief contains clear instructions and no ambiguity exists, proceed without asking. Zero-instruction processing is a core feature: the brief itself is the instruction set.

当出现以下任何一种情况时，请向用户请求澄清：

模糊的简报 — 研究目标不明确，存在多种解读
模糊的搜索指令 — 网络研究指令不清晰
知识不足 — 简报涉及你无法可靠处理的领域
数据矛盾 — 笔记中包含冲突信息且无明确解决方案
范围模糊 — 不清楚用户想要摘要、分析、新内容还是其他输出

当简报包含明确指令且无歧义时，无需询问即可直接处理。零指令处理是核心功能：简报本身就是指令集。

Step 5: Execute and Package

步骤5：执行与打包

Execute the work based on the brief content and any user instructions. Delegate to other installed skills as needed (e.g., web search, image generation, document creation).

When producing notes as output, ALWAYS use this mandatory header format:

markdown

undefined

根据简报内容和用户指令执行工作。必要时委托给其他已安装的技能（例如网络搜索、图片生成、文档创建）。

当生成笔记作为输出时，必须使用以下强制标题格式：

markdown

undefined

note title

笔记标题

short description

note content here...


The `# title` is the note's display name. The `> description` is a one-line summary. The `---` separator always precedes body content.

After completing the work, ALWAYS use the **athena-package** skill to package results for import back into Athena. Pass all produced notes, relevant assets, and appropriate aurora tags to the packaging step.

简短描述

笔记内容...


`# 标题`是笔记的显示名称。`> 描述`是一行摘要。`---`分隔符必须位于正文内容之前。

完成工作后，**必须**使用**athena-package**技能将结果打包，以便重新导入Athena。将所有生成的笔记、相关资源和适当的aurora标签传递给打包步骤。

Metadata Precedence

元数据优先级

When the same data appears in multiple files:

manifest.json is authoritative for IDs, paths, and scores
summaries.json is authoritative for note summaries
YAML frontmatter in references is convenience — use when reading individual notes
brief.md tables are human-readable summaries — use for quick orientation

当同一数据出现在多个文件中时：

manifest.json 是权威来源，用于ID、路径和评分
summaries.json 是权威来源，用于笔记摘要
参考笔记中的YAML前置内容是便捷信息——阅读单个笔记时使用
brief.md 表格是人类可读的摘要——用于快速了解内容

Multi-Brief Workflows

多简报工作流

When the user uploads multiple

.athenabrief

packages:

Process each brief independently first
Look for cross-brief connections using
```
knowledgeEntries
```
(shared entities, themes, linked notes)
Synthesize findings across briefs when the user asks for combined analysis
Package all results into a single
```
.athena
```
package unless the user requests separate packages

当用户上传多个

.athenabrief

包时：

首先独立处理每个简报
使用
```
knowledgeEntries
```
寻找跨简报的关联（共享实体、主题、链接的笔记）
当用户要求合并分析时，综合所有简报的发现
除非用户要求单独打包，否则将所有结果打包为单个
```
.athena
```
包

Delegation Patterns

委托模式

The brief may require capabilities beyond text processing:

Brief content	Delegate to
Web research instructions present	Web search tools
"Create a presentation" in objective	Presentation skills if available
"Generate images" in objective	Image generation skills if available
Code-related analysis	Code analysis tools
No delegation needed	Process directly

Always check what skills and tools are available before promising delegation.

简报可能需要文本处理之外的能力：

简报内容	委托给
包含网络研究指令	网络搜索工具
目标中包含“创建演示文稿”	演示文稿技能（如果可用）
目标中包含“生成图片”	图片生成技能（如果可用）
代码相关分析	代码分析工具
无需委托	直接处理

在承诺委托之前，请务必检查可用的技能和工具。

Error Handling

错误处理

Error	Recovery
ZIP extraction fails	Report the error; ask user to re-export from Athena
`instructions.md` missing	Proceed with brief.md; the reading order is documented here
`brief.md` missing	Fall back to summaries.json for orientation; warn user
`summaries.json` malformed	Fall back to manifest.json + brief.md table
`manifest.json` malformed	Use brief.md table + read references directly
Reference note missing	Skip it; warn user; use summary data instead
Asset file missing	Note the broken reference; continue processing
All core files missing	Report package appears corrupt; ask for re-export

错误	恢复方式
ZIP提取失败	报告错误；请用户从Athena重新导出
`instructions.md` 缺失	继续使用 `brief.md` ；阅读顺序已在此处记录
`brief.md` 缺失	退回到 `summaries.json` 了解内容；向用户发出警告
`summaries.json` 格式错误	退回到 `manifest.json` + `brief.md` 表格
`manifest.json` 格式错误	使用 `brief.md` 表格 + 直接阅读参考笔记
参考笔记缺失	跳过该笔记；向用户发出警告；改用摘要数据
资源文件缺失	记录损坏的引用；继续处理
所有核心文件缺失	报告包似乎已损坏；请用户重新导出

Graceful Degradation Hierarchy

优雅降级层级

If files are missing or malformed, fall back in this order:

```
brief.md
```
+
```
summaries.json
```
(ideal path)
```
brief.md
```
+
```
manifest.json
```
(summaries unavailable)
```
brief.md
```
alone (both JSON files broken)
```
manifest.json
```
+
```
references/*.md
```
(brief missing)
```
references/*.md
```
scanned directly (minimal package)

At each level, warn the user about missing data and reduced accuracy.

如果文件缺失或格式错误，按以下顺序降级处理：

```
brief.md
```
+
```
summaries.json
```
（理想路径）
```
brief.md
```
+
```
manifest.json
```
（摘要不可用）
仅
```
brief.md
```
（两个JSON文件均损坏）
```
manifest.json
```
+
```
references/*.md
```
（简报缺失）
直接扫描
```
references/*.md
```
（最简包）

在每个层级，都要向用户警告数据缺失和准确性降低的情况。

Example

示例

Input: User uploads

q1-strategy.athenabrief

Processing:

Extract ZIP, read
```
instructions.md
```
(standard reading guide)
Read
```
brief.md
```
— Research objective: "Review Q1 product strategy and roadmap priorities"
Read
```
summaries.json
```
— 5 notes covering roadmap, standups, competitor analysis, sprint retros, launch checklist
Short-circuit: User said "summarize this brief" — brief.md + summaries.json sufficient
Produce executive summary note + key findings note + action items note

Output: Hand off 3 notes to athena-package skill:

"Q1 Strategy Summary" (aurora:
```
focus
```
)
"Key Findings" (aurora:
```
explore
```
)
"Action Items" (aurora:
```
commitments
```
)

输入： 用户上传

q1-strategy.athenabrief

处理过程：

提取ZIP包，阅读
```
instructions.md
```
（标准阅读指南）
阅读
```
brief.md
```
— 研究目标：“回顾Q1产品战略和路线图优先级”
阅读
```
summaries.json
```
— 5篇笔记，涵盖路线图、站会、竞品分析、 sprint回顾、发布清单
短路判断：用户要求“总结这个简报” —
```
brief.md
```
+
```
summaries.json
```
已足够
生成执行摘要笔记 + 关键发现笔记 + 行动项笔记

输出： 将3篇笔记交给athena-package技能：

"Q1战略摘要"（aurora:
```
focus
```
）
"关键发现"（aurora:
```
explore
```
）
"行动项"（aurora:
```
commitments
```
）

Reference Materials

参考资料

references/athenabrief-spec.md
— Full
```
.athenabrief
```
format specification: ZIP structure, brief.md sections, summaries.json schema, manifest.json schema, reference note YAML frontmatter, asset naming conventions. Read when encountering unusual fields, debugging malformed packages, or building custom extraction logic.
references/processing-strategies.md
— Edge case handling, token optimization strategies, graceful degradation details, and advanced multi-brief workflows. Read when facing complex briefs, performance concerns, or unusual package configurations.

references/athenabrief-spec.md
— 完整的
```
.athenabrief
```
格式规范：ZIP结构、
```
brief.md
```
章节、
```
summaries.json
```
schema、
```
manifest.json
```
schema、参考笔记YAML前置内容、资源命名约定。遇到不常见字段、调试格式错误的包或构建自定义提取逻辑时阅读。
references/processing-strategies.md
— 边缘情况处理、令牌优化策略、优雅降级细节以及高级多简报工作流。处理复杂简报、性能问题或不常见包配置时阅读。