resume-intelligence

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Resume Intelligence

简历智能系统

Use this skill to convert approved career evidence into a professional resume and interview system: source inventory, tool usage log, evidence log, draft resume, detailed interview resume, career master run, career knowledge vault, career positioning coach, ATS/recruiter scorecard, polished resume, optional ATS-safe DOCX, optional designed HTML/PDF/Figma/Canva-ready resume, optional designed DOCX companion, optional targeted resume, optional cover letter, job/market fit analysis, job match scoring report, project interview briefs, technical-stack interview guide, interview readiness scorecard, interview prep pack, disclosure approval queue, redaction review, and LinkedIn profile recommendations.

使用本技能将已确认的职业经历转化为专业简历及面试相关系统内容：来源清单、工具使用日志、经历记录、简历草稿、详细面试简历、职业规划全流程文档、职业知识库、职业定位指导、ATS/招聘者评分卡、优化后的简历、可选的ATS兼容DOCX文件、可选的适配HTML/PDF/Figma/Canva的设计版简历、可选的配套设计版DOCX文件、可选的针对性简历、可选的求职信、职位/市场适配分析、职位匹配评分报告、项目面试简报、技术栈面试指南、面试准备就绪评分卡、面试准备包、披露审批队列、内容编辑审查、LinkedIn资料优化建议。

Guardrails

防护规则

Do not scrape LinkedIn behind login or bypass platform controls.
Use LinkedIn profile exports, pasted profile content, public profile content, or user notes.
Do not store credentials in plugin files, generated outputs, templates, or evidence logs.
Do not upload profile pictures, private documents, or resume content to Canva, Figma, or another third-party design service automatically.
Treat GitHub Enterprise, private Confluence, private Jira, and private repositories as confidential by default.
Preserve source provenance for every claim.
Mark uncertain claims for user confirmation.
Ask before including internal project names, customer names, repository names, Jira IDs, Confluence titles, private URLs, or exact internal metrics in final user-facing outputs.
Do not copy wording from other professionals' public profiles.

不得在登录后爬取LinkedIn内容或绕过平台管控。
使用LinkedIn资料导出文件、粘贴的资料内容、公开资料内容或用户笔记。
不得在插件文件、生成输出、模板或经历记录中存储凭证信息。
不得自动将头像、私人文档或简历内容上传至Canva、Figma或其他第三方设计服务。
默认将GitHub Enterprise、私有Confluence、私有Jira及私有仓库视为机密内容。
保留每项内容的来源出处。
对不确定的内容标记为需用户确认。
在最终面向用户的输出中包含内部项目名称、客户名称、仓库名称、Jira ID、Confluence标题、私有URL或精确内部指标前，需先询问用户。
不得复制其他专业人士公开资料中的表述。

When To Load References

何时加载参考文档

Read
```
references/source-model.md
```
before collecting evidence from more than one source or any GitHub source.
Read
```
references/evidence-schema.md
```
before creating structured evidence, run summaries, or quality-gate checks.
Read
```
references/privacy-sanitization.md
```
before writing final resume, cover letter, or LinkedIn text from enterprise or private sources.
Read
```
references/enterprise-redaction-audit.md
```
before finalizing outputs from GitHub Enterprise, private Confluence, Jira, private repositories, or confidential local documents.
Read
```
references/resume-workflow.md
```
before producing the evidence log, resume files, cover letter, or LinkedIn recommendations.
Read
```
references/career-coaching-workflow.md
```
before producing career master runs, career positioning advice, interview coaching, readiness scoring, or missing-information loops.
Read
```
references/career-vault-workflow.md
```
before creating or querying a durable Career Knowledge Vault, Obsidian vault, LLM wiki, or future-use career memory.
Read
```
references/docx-template-workflow.md
```
before producing DOCX, Word, Google Docs-targeted, PDF-export, Canva-ready, Figma-ready, or visually polished resume artifacts.
Read
```
references/visual-design-tools-workflow.md
```
before producing visually rich HTML/PDF output, Canva-ready packages, Figma handoff, browser-rendered screenshots, or deciding whether DOCX is the right final visual surface.
Read
```
references/visual-template-profiles.md
```
before selecting a named visual resume profile or using a profile picture in a visual artifact.
Read
```
references/interview-job-workflow.md
```
before producing project interview briefs, technical-stack interview guidance, interview preparation content, job search results, or job match scoring.
Read
```
references/open-source-onboarding.md
```
and
```
references/prompt-library.md
```
when the user asks for setup help, local-only mode, open-source installation, or reusable prompts.

从多个来源或任何GitHub来源收集经历前，阅读
```
references/source-model.md
```
。
创建结构化经历、流程总结或质量门检查前，阅读
```
references/evidence-schema.md
```
。
从企业或私有来源撰写最终简历、求职信或LinkedIn文本前，阅读
```
references/privacy-sanitization.md
```
。
从GitHub Enterprise、私有Confluence、Jira、私有仓库或机密本地文档生成最终输出前，阅读
```
references/enterprise-redaction-audit.md
```
。
生成经历记录、简历文件、求职信或LinkedIn建议前，阅读
```
references/resume-workflow.md
```
。
生成职业规划全流程文档、职业定位建议、面试指导、就绪评分或信息补全环节前，阅读
```
references/career-coaching-workflow.md
```
。
创建或查询持久化职业知识库、Obsidian库、LLM wiki或未来可用的职业记忆库前，阅读
```
references/career-vault-workflow.md
```
。
生成DOCX、Word、Google Docs适配版、PDF导出版、Canva适配版、Figma适配版或可视化优化简历成果前，阅读
```
references/docx-template-workflow.md
```
。
生成视觉丰富的HTML/PDF输出、Canva适配包、Figma交付文件、浏览器渲染截图，或决定DOCX是否为合适的最终视觉载体前，阅读
```
references/visual-design-tools-workflow.md
```
。
选择指定的可视化简历模板或在视觉成果中使用头像前，阅读
```
references/visual-template-profiles.md
```
。
生成项目面试简报、技术栈面试指导、面试准备内容、求职结果或职位匹配评分前，阅读
```
references/interview-job-workflow.md
```
。
当用户询问设置帮助、本地专属模式、开源安装或可复用提示词时，阅读
```
references/open-source-onboarding.md
```
和
```
references/prompt-library.md
```
。

Intake

需求收集

Start by identifying the run goal:

General resume.
Detailed interview resume or master career timeline.
Role-specific resume.
Company-specific resume.
Job-posting-specific resume and cover letter.
LinkedIn profile recommendations.
Career positioning or interview coaching.
Career Knowledge Vault, Obsidian vault, LLM wiki, or future-use knowledge base.
Professional DOCX, HTML/PDF, Figma-ready, Canva-ready, or visual template generation.
Interview preparation from resume evidence.
Job search and match scoring.
Evidence extraction only.
Safe local-only run or setup validation.

Collect only the inputs needed for that goal:

Years of experience and target seniority.
Target roles, technologies, domains, and geography.
LinkedIn profile export, pasted profile, public profile content, or notes.
Local project-document folder path.
Confluence CQL queries, spaces, titles, or page URLs approved by the user.
Jira JQL queries, projects, issue keys, assignee identity, and date ranges approved by the user.
Public GitHub account, repositories, PRs, issues, and date ranges.
GitHub Enterprise host, username, organization, repositories, credential context, contribution type, date range, and disclosure level.
Job posting text, URL content, or pasted listing.
Optional local profile-picture path or attachment for designed resume templates.
Confidentiality and redaction preferences.
Safe local-only mode preference and any third-party upload restrictions.

首先确定流程目标：

通用简历。
详细面试简历或完整职业时间线。
特定岗位简历。
特定公司简历。
适配特定职位招聘信息的简历及求职信。
LinkedIn资料优化建议。
职业定位或面试指导。
职业知识库、Obsidian库、LLM wiki或未来可用知识库。
专业DOCX、HTML/PDF、Figma适配版、Canva适配版或可视化模板生成。
基于简历经历的面试准备。
求职及匹配评分。
仅提取经历信息。
安全的本地专属流程或设置验证。

仅收集实现该目标所需的输入：

工作经验年限及目标职级。
目标岗位、技术栈、领域及地域。
LinkedIn资料导出文件、粘贴的资料、公开资料内容或笔记。
本地项目文档文件夹路径。
用户批准的Confluence CQL查询、空间、标题或页面URL。
用户批准的Jira JQL查询、项目、问题编号、经办人身份及日期范围。
公开GitHub账号、仓库、PR、问题及日期范围。
GitHub Enterprise主机、用户名、组织、仓库、凭证上下文、贡献类型、日期范围及披露级别。
职位招聘文本、URL内容或粘贴的招聘信息。
可选的用于设计版简历模板的本地头像路径或附件。
保密及内容编辑偏好。
安全本地专属模式偏好及任何第三方上传限制。

Source Collection

来源收集

Create a source context for each approved source. Keep public GitHub and GitHub Enterprise separate even when the same person owns both identities.

Always create

source-inventory.md

and

tool-usage-log.md

before

evidence.md

. The inventory must prove what was in scope, what was scanned, what was skipped, and why. The tool log must record the concrete tools, connector calls, shell commands, browser actions, or manual inputs used for each source. Do not silently infer that a folder or connector was read.

For local document folders, recursively enumerate the approved root with

rg --files <folder>

whenever available. Use a

find

fallback only when

rg

is unavailable or unsuitable. Do not stop after the first subfolder. Include every nested file in the inventory, then classify files as scanned, skipped-unsupported-format, skipped-too-large, skipped-binary, skipped-permission, duplicate, or needs-user-input. Prefer text-oriented formats such as

.md

.txt

.csv

.json

.yaml

.yml

.html

.xml

.adoc

.rst

.log

.docx

, and

.pdf

when local extraction tools are available; record extraction limitations for formats that cannot be parsed.

If connector tools are available, use them after the user approves the query or repository scope:

GitHub: repositories, authored pull requests, reviewed pull requests, issues, commits, comments, and diffs when relevant.
Confluence: user-approved CQL searches and specific pages.
Jira: user-approved JQL searches and selected fields.
Local files: inspect text-oriented documents under the approved folder.
Web research: use current public sources for job postings and public market research when the user asks for optimization or targeting.
Profile picture: use only a user-provided local image path or attachment, and include it only in designed artifacts when requested.

If a connector is unavailable, credentials are missing, or a scoped source cannot be accessed, record the attempt in

tool-usage-log.md

, mark the source context as unavailable or needs-user-input, ask for pasted exports, local files, or manual summaries, and continue with other approved sources.

为每个已批准的来源创建来源上下文。即使同一人拥有公开GitHub和GitHub Enterprise账号，也要将两者分开处理。

在创建

evidence.md

前，务必先创建

source-inventory.md

和

tool-usage-log.md

。来源清单必须证明纳入范围的内容、已扫描的内容、已跳过的内容及原因。工具日志必须记录每个来源所使用的具体工具、连接器调用、Shell命令、浏览器操作或手动输入。不得默认推断已读取某个文件夹或连接器。

对于本地文档文件夹，只要可用，就使用

rg --files <folder>

递归枚举已批准的根目录。仅当

rg

不可用或不适用时，才使用

find

作为备选。不得在第一个子文件夹后停止扫描。将每个嵌套文件纳入清单，然后将文件分类为已扫描、因格式不支持跳过、因文件过大跳过、因二进制文件跳过、因权限不足跳过、重复文件或需用户输入。当本地提取工具可用时，优先处理面向文本的格式，如

.md

、

.txt

、

.csv

、

.json

、

.yaml

、

.yml

、

.html

、

.xml

、

.adoc

、

.rst

、

.log

、

.docx

和

.pdf

；对于无法解析的格式，记录提取限制。

如果连接器工具可用，在用户批准查询或仓库范围后使用：

GitHub：仓库、作者提交的拉取请求、审核的拉取请求、问题、提交记录、评论及相关差异。
Confluence：用户批准的CQL搜索结果及特定页面。
Jira：用户批准的JQL搜索结果及选定字段。
本地文件：检查已批准文件夹下的面向文本文档。
网络调研：当用户要求优化或针对性内容时，使用当前公开来源获取职位招聘信息及公开市场调研数据。
头像：仅使用用户提供的本地图片路径或附件，且仅在用户要求时纳入设计版成果。

如果连接器不可用、凭证缺失或无法访问指定范围的来源，在

tool-usage-log.md

中记录尝试情况，将来源上下文标记为不可用或需用户输入，请求用户提供粘贴的导出文件、本地文件或手动摘要，然后继续处理其他已批准的来源。

Evidence Extraction

经历提取

For each source, extract:

Project or product name.
Source context and evidence reference.
User role and ownership.
Business, customer, technical, operational, security, or platform problem.
Actions performed by the user.
Technologies and platforms.
Scale, reliability, performance, cost, adoption, security, migration, or operational impact.
Dates or time windows.
Quantitative metrics.
Collaborators and cross-functional scope.
Confidence level: high, medium, or low.
Disclosure level: public, internal-summary-only, confidential, excluded, or unknown.

Prefer direct evidence over inference. When evidence conflicts, ask the user which version is accurate.

从每个来源提取以下信息：

项目或产品名称。
来源上下文及经历引用。
用户角色及职责归属。
业务、客户、技术、运营、安全或平台问题。
用户执行的操作。
技术栈及平台。
规模、可靠性、性能、成本、采用率、安全性、迁移或运营影响。
日期或时间范围。
量化指标。
协作人员及跨职能范围。
置信度：高、中或低。
披露级别：公开、仅内部摘要、机密、排除或未知。

优先使用直接证据而非推断内容。当证据存在冲突时，询问用户哪个版本准确。

Output Location

输出位置

Create outputs under:

text

resume-runs/YYYY-MM-DD-<descriptive-slug>/

Use these filenames when the corresponding output is requested:

```
source-inventory.md
```
```
tool-usage-log.md
```
```
evidence.md
```
```
run-summary.md
```
```
safe-local-mode-checklist.md
```
```
disclosure-approval-queue.md
```
```
final-redaction-review.md
```
```
draft-resume.md
```
```
detailed-resume.md
```
```
career-master-run.md
```
```
career-vault/
```
```
career-vault-lint.md
```
```
missing-information-loop.md
```
```
career-positioning-coach.md
```
```
resume-scorecard.md
```
```
professional-resume.md
```
```
ats-resume.docx
```
```
designed-resume.docx
```
```
designed-resume.html
```
```
designed-resume.pdf
```
```
visual-template-selector.md
```
```
visual-design-review.md
```
```
visual-tool-handoff.md
```
```
targeted-resume.md
```
```
cover-letter.md
```
```
linkedin-recommendations.md
```
```
gap-analysis.md
```
```
job-match-report.md
```
```
project-interview-briefs.md
```
```
technical-stack-interview-guide.md
```
```
interview-readiness-scorecard.md
```
```
interview-prep-pack.md
```
```
docx-render-review.md
```

Use the templates in

assets/templates/

as structure, not as rigid wording.

在以下路径下创建输出：

text

resume-runs/YYYY-MM-DD-<descriptive-slug>/

当请求对应的输出时，使用以下文件名：

```
source-inventory.md
```
```
tool-usage-log.md
```
```
evidence.md
```
```
run-summary.md
```
```
safe-local-mode-checklist.md
```
```
disclosure-approval-queue.md
```
```
final-redaction-review.md
```
```
draft-resume.md
```
```
detailed-resume.md
```
```
career-master-run.md
```
```
career-vault/
```
```
career-vault-lint.md
```
```
missing-information-loop.md
```
```
career-positioning-coach.md
```
```
resume-scorecard.md
```
```
professional-resume.md
```
```
ats-resume.docx
```
```
designed-resume.docx
```
```
designed-resume.html
```
```
designed-resume.pdf
```
```
visual-template-selector.md
```
```
visual-design-review.md
```
```
visual-tool-handoff.md
```
```
targeted-resume.md
```
```
cover-letter.md
```
```
linkedin-recommendations.md
```
```
gap-analysis.md
```
```
job-match-report.md
```
```
project-interview-briefs.md
```
```
technical-stack-interview-guide.md
```
```
interview-readiness-scorecard.md
```
```
interview-prep-pack.md
```
```
docx-render-review.md
```

将

assets/templates/

中的模板用作结构参考，而非固定表述。

Output Sequence

输出流程

Build
```
source-inventory.md
```
and
```
tool-usage-log.md
```
first.
If the user wants local-only mode, create
```
safe-local-mode-checklist.md
```
and avoid connector, browser, Figma, Canva, and job-board access unless the user later approves it.
Build
```
evidence.md
```
from scanned or manually provided sources only.
If private or enterprise sources are used, create
```
disclosure-approval-queue.md
```
before final user-facing wording.
Create
```
draft-resume.md
```
from high- and medium-confidence evidence.
If the user wants full interview preparation, career reconstruction, "all work timeline" detail, or career coaching, create
```
detailed-resume.md
```
,
```
career-master-run.md
```
, and
```
missing-information-loop.md
```
before the scorecard. Include full work chronology, roles, responsibilities, project ownership, impact, metrics, technologies, leadership scope, and questions for missing information.
If the user wants future optimization, durable knowledge, Obsidian, an LLM wiki, or reusable career memory, create or update
```
career-vault/
```
from structured evidence, then run
```
scripts/vault_lint.py
```
and save the result as
```
career-vault-lint.md
```
.
Ask the user to confirm low-confidence claims and sensitive details before final wording.
Create
```
resume-scorecard.md
```
before final wording. Check page-length strategy, ATS parse safety, first-page recruiter strength, keyword coverage, unsupported claims, and job-posting gaps.
If career positioning is requested or useful for a senior/profile rewrite, create
```
career-positioning-coach.md
```
from the detailed resume and scorecard.
Create
```
professional-resume.md
```
using sanitized, evidence-backed bullets and the scorecard recommendations.
If DOCX, Word, or Google Docs is requested, create
```
ats-resume.docx
```
as the primary application version and optionally
```
designed-resume.docx
```
for recruiter/referral sharing. Use
```
scripts/build_resume_docx.py
```
when it fits the content, then render-check the DOCX with the Documents skill before delivery.
If an impressive visual resume, recruiter-facing resume, Canva/Figma-ready layout, profile picture, or PDF export is requested, create
```
visual-template-selector.md
```
, make
```
designed-resume.html
```
the primary visual artifact, render it with browser tooling when available, and create
```
designed-resume.pdf
```
,
```
visual-design-review.md
```
, and
```
visual-tool-handoff.md
```
when requested or useful. Use Figma tools only when the user provides or approves a Figma target file; use Canva only as a manual/import handoff unless the user explicitly approves uploading sanitized content.

If a job posting is provided, create

gap-analysis.md

targeted-resume.md

, and

cover-letter.md

If job search is requested, search current public job sources and create
```
job-match-report.md
```
with match scoring and citations.

If interview preparation is requested, create

project-interview-briefs.md

technical-stack-interview-guide.md

interview-prep-pack.md

, and

interview-readiness-scorecard.md

Create
```
linkedin-recommendations.md
```
with editable profile updates.
If private or enterprise sources informed final outputs, create
```
final-redaction-review.md
```
before delivery.
Create
```
run-summary.md
```
for substantial runs summarizing source coverage, tools used, outputs created, warning-only quality gates, sensitive approval needs, career vault status, and missing user questions.

首先创建
```
source-inventory.md
```
和
```
tool-usage-log.md
```
。
如果用户需要本地专属模式，创建
```
safe-local-mode-checklist.md
```
，除非用户后续批准，否则避免使用连接器、浏览器、Figma、Canva及求职板访问。
仅从已扫描或手动提供的来源创建
```
evidence.md
```
。
如果使用了私有或企业来源，在生成最终面向用户的表述前，创建
```
disclosure-approval-queue.md
```
。
从高置信度和中置信度的经历中创建
```
draft-resume.md
```
。
如果用户需要完整的面试准备、职业经历重构、“全工作时间线”细节或职业教练指导，在生成评分卡前创建
```
detailed-resume.md
```
、
```
career-master-run.md
```
和
```
missing-information-loop.md
```
。包含完整的工作时间线、角色、职责、项目归属、影响、指标、技术栈、领导力范围及需补充信息的问题。
如果用户需要未来优化、持久化知识、Obsidian库、LLM wiki或可复用职业记忆库，从结构化经历中创建或更新
```
career-vault/
```
，然后运行
```
scripts/vault_lint.py
```
并将结果保存为
```
career-vault-lint.md
```
。
在生成最终表述前，请求用户确认低置信度内容及敏感细节。
在生成最终表述前创建
```
resume-scorecard.md
```
。检查页面长度策略、ATS解析安全性、首页对招聘者的吸引力、关键词覆盖率、无依据内容及与职位招聘信息的差距。
如果用户请求职业定位或该内容对资深/个人资料重写有用，从详细简历和评分卡中创建
```
career-positioning-coach.md
```
。
使用已脱敏、有经历支撑的要点及评分卡建议创建
```
professional-resume.md
```
。
如果用户请求DOCX、Word或Google Docs版本，创建
```
ats-resume.docx
```
作为主要申请版本，可选创建
```
designed-resume.docx
```
用于招聘者/推荐人分享。当内容适配时使用
```
scripts/build_resume_docx.py
```
，然后在交付前使用文档技能对DOCX进行渲染检查。
如果用户请求令人印象深刻的可视化简历、面向招聘者的简历、适配Canva/Figma的布局、头像或PDF导出，创建
```
visual-template-selector.md
```
，将
```
designed-resume.html
```
作为主要视觉成果，若可用则使用浏览器工具渲染，并在请求或需要时创建
```
designed-resume.pdf
```
、
```
visual-design-review.md
```
和
```
visual-tool-handoff.md
```
。仅当用户提供或批准Figma目标文件时使用Figma工具；仅在用户明确批准上传脱敏内容时，才可将Canva作为手动/导入交付方式。
如果提供了职位招聘信息，创建
```
gap-analysis.md
```
、
```
targeted-resume.md
```
和
```
cover-letter.md
```
。
如果用户请求求职服务，搜索当前公开求职来源并创建
```
job-match-report.md
```
，包含匹配评分及引用来源。

如果用户请求面试准备，创建

project-interview-briefs.md

、

technical-stack-interview-guide.md

、

interview-prep-pack.md

和

interview-readiness-scorecard.md

。

创建包含可编辑资料更新建议的
```
linkedin-recommendations.md
```
。
如果私有或企业来源为最终输出提供了信息，在交付前创建
```
final-redaction-review.md
```
。
对于重要流程，创建
```
run-summary.md
```
，总结来源覆盖范围、使用的工具、生成的输出、仅警告的质量门、需敏感内容批准的事项、职业知识库状态及需用户补充的问题。

Final Response

最终回复

Summarize:

Files created.
Sources used and skipped.
Inventory and tool-log coverage, including any approved folders or connector scopes that were unavailable.
Sensitive items that require user review.
Low-confidence claims that need confirmation.
Detailed resume gaps: missing dates, responsibilities, metrics, scope, project outcomes, or stories that require user input.
Career positioning and interview readiness gaps when coaching outputs were requested.
Career vault path, lint status, and any missing evidence links when durable knowledge was requested.
Scorecard findings for ATS safety, recruiter strength, page length, and keyword gaps.
Visual tool path used: HTML/PDF, Figma, Canva handoff, DOCX, or multiple outputs.
Visual template profile and profile-picture approval status when visual outputs were requested.
Disclosure approval and final redaction status when private sources were used.
DOCX template used, render-check status, and whether the delivered DOCX is ATS-safe, designed, or both.
Strongest job matches and interview preparation gaps when those outputs were requested.
Clear next actions for the user.

总结以下内容：

创建的文件。
使用和跳过的来源。
来源清单和工具日志的覆盖范围，包括任何不可用的已批准文件夹或连接器范围。
需要用户审核的敏感内容。
需要确认的低置信度内容。
详细简历的差距：缺失的日期、职责、指标、范围、项目成果或需用户补充的故事。
当请求指导输出时，职业定位和面试准备的差距。
当请求持久化知识时，职业知识库路径、lint状态及任何缺失的经历链接。
评分卡关于ATS安全性、招聘者吸引力、页面长度及关键词差距的发现。
使用的视觉工具路径：HTML/PDF、Figma、Canva交付、DOCX或多种输出。
当请求视觉输出时，可视化模板配置文件及头像批准状态。
当使用私有来源时，披露批准及最终内容编辑状态。
使用的DOCX模板、渲染检查状态，以及交付的DOCX是否为ATS兼容版、设计版或两者兼具。
当请求相关输出时，最匹配的职位及面试准备差距。
明确的用户下一步行动。