hubspot

作者: openai

用于销售工作流的HubSpot CRM连接器指南,支持读取、草稿、备注或建议记录变更。

npx skills add https://github.com/openai/role-specific-plugins --skill hubspot

HubSpot User Guide

Prepare trustworthy HubSpot-backed context or a safely reviewed HubSpot action while keeping the surrounding Sales workflow authoritative for the business task.

Common Skill Instructions

MANDATORY: If not already in context, read and adhere closely to plugins/sales/skills/index/SKILL.md## Cross-Skill Best Practices.

HubSpot Operation Ranking

Use this priority order when the requested HubSpot job or record is ambiguous.

  1. Explicit user request, named record, HubSpot ID, company, deal, contact, ticket, or stated action
  2. Read-only CRM context needed by the active Sales workflow
  3. Exact record or a high-confidence candidate resolved from the user's words and business context
  4. The smallest property set, associations, activity context, or page of results that answers the request
  5. A reviewed draft of proposed HubSpot changes
  6. An explicitly approved, supported HubSpot write followed by verification

Do not silently pick among plausible records, broaden a read into a write, or treat clarification as write approval.

Key Dependency Categories

These are particularly important for this workflow; use your best judgment to potentially include other data sources to improve quality.

  • ~~CRM, specifically the exposed HubSpot connector, for HubSpot-backed identity, records, properties, associations, links, and approved writes
  • The parent Sales workflow for Calendar, Meeting Transcripts, Email, Internal Messaging, Knowledge & Files, or Sales Intelligence context that HubSpot does not own
  • User-provided record IDs, names, exports, property values, and business context when connector access is unavailable or a record needs disambiguation

If HubSpot is missing, unavailable, stale, ambiguous, or limited by the exposed connector surface, say so and label the answer as not HubSpot-backed.

Workflow Guidance

These HubSpot-specific steps override the default workflow in the index skill.

    1. Resolve access and target
    • Use get_user_details first to confirm the connected portal and exposed object read/write availability.
    • Resolve whether the user wants a read, a draft, or an approved write; identify object type, owner or team, pipeline, timeframe, stage, and exact record when those affect the answer.
    • For an ambiguous record, do the smallest bounded candidate lookup and ask the user to choose before using it for a summary or write.
    1. Discover properties and records
    • Use search_properties with at most five focused keywords when fields are uncertain; use get_properties to confirm enum values and writable properties.
    • Use search_crm_objects for search, counts, filters, pagination, and associations. Use get_crm_objects for known IDs. Do not use deprecated search or fetch.
    • Retrieve only the properties and associations that affect the sales decision. State filters, totals, page or sample limits, and whether analysis is sampled.
    1. Draft, approve, and verify
    • For proposed changes, show an exact field-level draft before calling manage_crm_objects.
    • Write only after the user explicitly approves the exact reviewed change or batch. Batch at most 10 objects and confirm associations separately.
    • A reviewed-batch approval applies only to that exact batch; never offer a blanket confirmation bypass.
    • Inspect the write response and verify with a narrow read when the response does not prove the outcome.

HubSpot Rules

  • Prefer exact IDs, domains, emails, and full names plus company over vague name-only matching.
  • Before querying or writing with a property or enum, confirm property labels, internal names, enum values, and writeability; do not guess them.
  • Keep HubSpot as CRM truth for HubSpot-owned account, contact, deal, ticket, owner, pipeline, and activity facts. Do not let enrichment or web context overwrite CRM truth.
  • Include clickable connector-returned or trusted HubSpot record URLs with UTM parameters when available.
  • Do not write inferred data, overwrite user-entered context without clear consent, or mutate an association without explicit approval.
  • Do not imply support for objects, properties, pagination behavior, bulk sizes, or writes that the live connector does not expose.

Connector Boundary

Use only the HubSpot surfaces that are actually exposed:

  • Identity and access: get_user_details
  • Property discovery: search_properties, get_properties
  • Record search and reads: search_crm_objects, get_crm_objects
  • Reviewed writes: manage_crm_objects

For unsupported broad exports, delete, merge, dedupe, arbitrary automation, or unexposed object operations, state the missing surface and offer a narrower read, a draft, or another approved source.

Modes

1. HubSpot Read

  • Use for company, deal, contact, ticket, owner, activity, property, or association context.
  • Resolve the target, retrieve only decision-relevant fields, and call out ambiguity or gaps.

Output Format

# [HubSpot Record Or Context]

## Summary

- [Most important CRM fact with trusted link]
- [Current deal, company, contact, activity, or association signal]
- [Risk, missing property, ambiguity, or source gap]

---

{Follow the instructions and output format/conditions in [Limitations and Improvements](../index/SKILL.md#limitations-and-improvements)}

{Follow the instructions and output format/conditions in [Next Steps](../index/SKILL.md#4-next-steps)}

2. HubSpot Update Draft

  • Use when the user asks to prepare a note, property change, association, create, or update but has not approved the write.
  • Keep current and proposed values separate and identify the exact target.

Output Format

# Proposed HubSpot Update

**Target:** [Record name, type, and trusted link or ID]

| Property / Association | Current Value | Proposed Value | Reason / Source |
| --- | --- | --- | --- |
| [Label] | [Current or Unknown] | [Proposed] | [Grounded reason + link] |

## Before I Write

- [Missing input, association confirmation, validation risk, or approval needed]

---

{Follow the instructions and output format/conditions in [Limitations and Improvements](../index/SKILL.md#limitations-and-improvements)}

{Follow the instructions and output format/conditions in [Next Steps](../index/SKILL.md#4-next-steps)}

3. Approved HubSpot Write

  • Use only after explicit approval for the exact reviewed object or batch.
  • Verify before reporting completion when the write response is inconclusive.

Output Format

# HubSpot Updated

**Record:** [Name, type, and trusted link or ID]

- **Changed:** [Property or association]
- **Result:** [Verified outcome]
- **Not changed:** [Unsupported or skipped item]

---

{Follow the instructions and output format/conditions in [Limitations and Improvements](../index/SKILL.md#limitations-and-improvements)}

{Follow the instructions and output format/conditions in [Next Steps](../index/SKILL.md#4-next-steps)}

来自 openai 的更多技能

user-context
openai
加载或管理数据分析插件的持久化源路由偏好、引导逻辑、设置进度及语义层注册表。
official
notion-research-documentation
openai
研究Notion内容,并将其综合成带有引用的结构化简报、报告或对比。通过定向查询搜索并获取Notion页面,然后按主题组织发现,附带内联来源引用和参考文献部分。根据范围和用户目标,从四种输出格式(快速简报、研究摘要、对比、综合报告)中选择。使用内置模板创建和更新Notion页面;直接链接来源,并在新信息到达时跟踪变更...
official
rcsb-pdb-skill
openai
提交紧凑的RCSB PDB请求以获取核心元数据、Search API查询和FASTA下载。当用户需要简洁的RCSB摘要时使用;保存原始JSON或…
official
pdf
openai
PDF的读取、创建与验证,支持可视化渲染与程序化生成。使用Poppler(pdftoppm)将PDF页面渲染为PNG,以便在交付前直观检查布局、间距与排版;通过reportlab程序化生成PDF,确保格式可靠;利用pdfplumber或pypdf提取文本与元数据。执行质量标准:无文本裁剪、元素重叠、表格损坏或渲染伪影;仅使用ASCII连字符,引用内容需可读。使用...
official
test-coverage-improver
openai
改进OpenAI Agents JS mon
official
playwright
openai
基于终端驱动的浏览器自动化,支持元素快照与交互式UI工作流。通过playwright-cli包装脚本运行(需npx),支持无头模式与有头模式进行可视化调试。核心工作流:打开页面、获取快照以稳定元素引用、使用引用进行交互、在导航或DOM变更后重新快照。包含表单填写、点击、输入、多标签页管理、截图/PDF捕获及用于流程调试的追踪记录。元素引用(如e3、e15)...
official
ukb-topmed-phewas-skill
openai
通过接受rsID、GRCh37或GRCh38输入并解析为所需的GRCh38查询,获取单个变体的紧凑型UKB-TOPMed PheWAS摘要。当需要…时使用。
official
code-review-context
openai
模型可见上下文
official