sales-agentforce-pipeline-management-configure

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Agentforce Pipeline Management Configuration

Agentforce Pipeline Management 配置

Configure Agentforce Pipeline Management end-to-end in a Salesforce org. This skill handles both greenfield orgs (zero configuration) and partially configured orgs that need repair or completion.

在Salesforce组织中端到端配置Agentforce Pipeline Management。本技能可处理全新组织（零配置）和需要修复或完成配置的部分配置组织。

Scope

适用范围

In scope: Enabling Pipeline Management, configuring the Sales Management agent, activating/customizing flows (Process Field Update Suggestions, Get Opportunity Grounding Data, Get AI Recommendations from Call Transcripts), prompt template configuration, permission set assignment, data source setup, opportunity stage descriptions, autonomous field update configuration, and partial-org repair.
Out of scope: Building custom agents from scratch (use
```
agentforce-generate
```
). Einstein Conversation Insights or Einstein Activity Capture initial provisioning (those are separate products). Slack app installation. Custom Apex triggers on Opportunity Team Member.

包含范围：启用Pipeline Management、配置Sales Management agent、激活/自定义Flow（Process Field Update Suggestions、Get Opportunity Grounding Data、Get AI Recommendations from Call Transcripts）、提示模板配置、Permission Set分配、数据源设置、Opportunity阶段描述、自动字段更新配置，以及部分组织修复。
排除范围：从头构建自定义agent（请使用
```
agentforce-generate
```
）。Einstein Conversation Insights或Einstein Activity Capture的初始资源调配（这些是独立产品）。Slack应用安装。Opportunity Team Member上的自定义Apex触发器。

Prerequisites

前提条件

You must verify (scripts cannot check or change these)

必须验证（脚本无法检查或修改这些内容）

Edition: Enterprise, Performance, Unlimited, or Developer Edition with Agentforce for Sales add-on (or Agentforce 1 Sales Edition). Setup probes this up front (Step 0) — it fails fast with a license/edition message on an org that can't run Pipeline Management, before asking anything — but edition/license provisioning itself is still admin-owned.

User permissions: The executing user needs

View Setup

AND (

Modify All Data

Customize Application

Manage AI Agents

AND (

Manage Agentforce Employee Agents

Customize Application

), and

Assign Permission Sets

版本：Enterprise、Performance、Unlimited或Developer版本，并带有Agentforce for Sales附加组件（或Agentforce 1 Sales Edition）。设置时会预先检查这一点（步骤0）——如果组织无法运行Pipeline Management，会快速返回许可证/版本错误信息，无需询问其他问题——但版本/许可证的资源调配仍由管理员负责。

用户权限：执行用户需要

View Setup

AND (

Modify All Data

Customize Application

)、

Manage AI Agents

AND (

Manage Agentforce Employee Agents

Customize Application

)，以及

Assign Permission Sets

权限。

Enabled automatically by

scripts/setup-all.sh

由

scripts/setup-all.sh

自动启用

The following settings are toggled by setup automatically — do not enable them manually before running the script.

Einstein Generative AI (
```
EinsteinGptSettings.enableEinsteinGptPlatform
```
) — required; must be on before Agentforce Agent.
Agentforce Agent (
```
EinsteinCopilotSettings.enableEinsteinGptCopilot
```
) — required; depends on Einstein Generative AI.
Agentforce Studio / Agent Platform (
```
AgentPlatformSettings.enableAgentPlatform
```
) — required; depends on Agentforce Agent and gates the Deal Agent. Core rejects
```
SalesDealAgentSettings.enableDealAgent
```
when this is off, surfacing as an opaque Deal Agent activation failure.
Enhanced Notes (
```
EnhancedNotesSettings.enableEnhancedNotes
```
) — required; Pipeline Management uses ContentNote as a primary data source.
Opportunity Team (
```
OpportunitySettings.enableOpportunityTeam
```
) — required; the suggestion flow uses OpportunityTeamMember and will fail deployment without it.
Pipeline Inspection (
```
OpportunitySettings.enablePipelineInspection
```
) — required; provides the UI where reps view and accept/dismiss suggestions. Suggestions still generate without it, but users have nowhere to see them.
Enhanced Email (
```
EmailAdministrationSettings.enableEnhancedEmailEnabled
```
) — optional but recommended; only needed for email-body indexing when Einstein Activity Capture is in use.
Pipeline Management (
```
SalesDealAgentSettings.enableDealAgent
```
) — required; the feature itself.

以下设置会由安装脚本自动开启——运行脚本前请勿手动启用。

Einstein Generative AI（
```
EinsteinGptSettings.enableEinsteinGptPlatform
```
）——必填；必须在启用Agentforce Agent之前开启。
Agentforce Agent（
```
EinsteinCopilotSettings.enableEinsteinGptCopilot
```
）——必填；依赖于Einstein Generative AI。
Agentforce Studio / Agent Platform（
```
AgentPlatformSettings.enableAgentPlatform
```
）——必填；依赖于Agentforce Agent，是Deal Agent的前置条件。如果此设置关闭，核心系统会拒绝
```
SalesDealAgentSettings.enableDealAgent
```
，表现为Deal Agent激活失败且无明确提示。
Enhanced Notes（
```
EnhancedNotesSettings.enableEnhancedNotes
```
）——必填；Pipeline Management将ContentNote作为主要数据源。
Opportunity Team（
```
OpportunitySettings.enableOpportunityTeam
```
）——必填；建议Flow会使用OpportunityTeamMember，若未启用则部署失败。
Pipeline Inspection（
```
OpportunitySettings.enablePipelineInspection
```
）——必填；提供销售代表查看、接受/驳回建议的UI。即使没有此设置，建议仍会生成，但用户无法查看。
Enhanced Email（
```
EmailAdministrationSettings.enableEnhancedEmailEnabled
```
）——可选但推荐；仅在使用Einstein Activity Capture时需要用于邮件正文索引。
Pipeline Management（
```
SalesDealAgentSettings.enableDealAgent
```
）——必填；功能本身。

Clarifying Questions

澄清问题

Run the license preflight FIRST — before asking any of these questions. As your very first action (right after resolving the org alias), run
bash scripts/setup-all.sh <org-alias> --check-license
. It runs only auth + the capability gate, asks nothing, and changes nothing. If it exits non-zero, relay the printed license/edition blocker to the user and STOP — do not ask any clarifying question and make no changes. Only when it exits 0 (org is capable — greenfield or already provisioned) proceed to the questions below.

Once the preflight passes, determine:

Greenfield or repair? Is this a new org that has never had Pipeline Management, or one with partial configuration?
Data sources: Which data sources should inform the agent?
- Notes (enabled by default)
- Emails (requires Einstein Activity Capture)
- Voice/Video calls (requires Einstein Conversation Insights)
- Enhanced Email (for email body indexing — requires
```
EmailAdministrationSettings.enableEnhancedEmailEnabled
```
  )
Autonomous updates: Should the agent update opportunity fields autonomously, or only suggest updates for user review?
Which fields should the agent manage? (ask this FIRST, up front — setup is field-selection-driven).
```
setup-all.sh
```
builds the flow with only the fields you select, from the start — there is no deploy-both-then-strip step. Ask: "Which Opportunity fields should the agent suggest values for? The OOTB options are Next Step (
NextStep
) and Opportunity Stage (
StageName
); you can also add custom text fields (e.g. a Competitor Analysis field)."
- At least 1 field is REQUIRED — setup aborts (exit 1) if zero fields are selected. There is no default to
```
NextStep
```
  .
- Hard cap: 5 fields total; OOTB fields count toward it. Selecting more aborts before any change.
- Pass the set with
```
--fields "NextStep,StageName,Risk__c"
```
  in non-interactive runs, or type it at the interactive prompt.
Per-field prompt customization (all fields EXCEPT
StageName
). For each selected non-
```
StageName
```
field, setup collects an optional goal ("You must think about…") and instruction (extraction guidance):
- NextStep
  ships with a curated OOTB managed prompt. Setup only overrides it when you supply a goal/instruction (interactive: it asks whether to customize; non-interactive: pass
```
--field-goal "NextStep:…"
```
  /
```
--field-instruction "NextStep:…"
```
  ). Skip customization to keep the OOTB prompt.
- Custom fields always get a goal/instruction (your text, or the script's sensible defaults).
- StageName
  is NOT customizable — it is a managed picklist template with no override support; setup never prompts for it.
Custom field eligibility: Only standard or custom text fields on Opportunity are supported — plain Text (
```
type=string
```
) or Text Area ≤ 255 (
```
type=textarea
```
,
```
htmlFormatted=false
```
); length ≤ 255; not Long Text Area (length > 255), Rich Text Area (
```
htmlFormatted=true
```
), picklist, or formula. The suggestion is capped at the field's own length. Fields selected in question 4 are wired during
```
setup-all.sh
```
; additional fields added after setup use
```
scripts/add-field-suggestion.sh <org-alias> <FieldApiName>
```
(one per field; see Phase 4.3). Both paths enforce the 5-total cap —
```
add-field-suggestion.sh
```
refuses to wire a 6th field.
Opportunity stages: Opportunity stages are picklist values, and different Opportunity record types may expose different stage subsets. Are standard or custom opportunity stages in use? Do their descriptions need to be defined/updated? Note:
```
OpptStageDescription
```
is global-per-stage — keyed only by
```
OpportunityStageApiName
```
, with no per-record-type column — so a description is written once per active stage and applies across every record type that exposes that stage. Setup works the same whether or not the org has Opportunity record types.

先运行许可证预检——在询问任何问题之前。 作为你的第一个操作（解析组织别名后立即执行），运行
bash scripts/setup-all.sh <org-alias> --check-license
。此命令仅执行身份验证和能力检查，不会询问任何问题，也不会修改任何内容。如果命令返回非零值，将打印的许可证/版本阻止信息告知用户并停止操作——不要询问任何澄清问题，也不要进行任何修改。只有当命令返回0（组织具备运行条件——全新或已完成资源调配）时，才继续询问以下问题。

预检通过后，确定：

全新组织还是修复？ 这是从未配置过Pipeline Management的新组织，还是部分配置的组织？
数据源：哪些数据源应为agent提供信息？
- Notes（默认启用）
- Emails（需要Einstein Activity Capture）
- 语音/视频通话（需要Einstein Conversation Insights）
- Enhanced Email（用于邮件正文索引——需要
```
EmailAdministrationSettings.enableEnhancedEmailEnabled
```
  ）
自动更新：agent应自动更新Opportunity字段，还是仅提供更新建议供用户审核？
agent应管理哪些字段？（请首先询问此问题——安装流程由字段选择驱动）
```
setup-all.sh
```
会从一开始就仅为你选择的字段构建Flow——不存在先部署所有字段再删除的步骤。询问：“agent应为哪些Opportunity字段建议值？默认选项是Next Step（
NextStep
）和Opportunity Stage（
StageName
）；你也可以添加自定义文本字段（例如竞争对手分析字段）。”
- 至少需要选择1个字段——如果未选择任何字段，安装会中止（退出码1）。没有默认选择
```
NextStep
```
  的设置。
- 上限：最多5个字段；默认字段计入总数。选择超过5个字段会在进行任何修改前中止安装。
- 在非交互式运行中，使用
```
--fields "NextStep,StageName,Risk__c"
```
  参数传递字段集，或在交互式提示中输入。
按字段自定义提示（除
StageName
外的所有字段）。对于每个选中的非
```
StageName
```
字段，安装流程会收集可选的目标（“你必须考虑……”）和指令（提取指导）：
- NextStep
  附带经过优化的默认托管提示。仅当你提供目标/指令时，安装流程才会覆盖默认提示（交互式：会询问是否自定义；非交互式：传递
```
--field-goal "NextStep:…"
```
  /
```
--field-instruction "NextStep:…"
```
  参数）。跳过自定义即可保留默认提示。
- 自定义字段始终需要目标/指令（可以是你输入的文本，或脚本的合理默认值）。
- StageName
  不可自定义——它是托管选择列表模板，不支持覆盖；安装流程永远不会针对此字段进行提示。
自定义字段 eligibility：仅支持Opportunity上的标准或自定义文本字段——纯文本（
```
type=string
```
）或长度≤255的文本区域（
```
type=textarea
```
，
```
htmlFormatted=false
```
）；长度≤255；不支持长文本区域（长度>255）、富文本区域（
```
htmlFormatted=true
```
）、选择列表或公式字段。建议内容长度受限于字段本身的长度。问题4中选择的字段会在
```
setup-all.sh
```
运行时关联；安装完成后添加的字段需使用
```
scripts/add-field-suggestion.sh <org-alias> <FieldApiName>
```
（每个字段运行一次；见第4.3阶段）。两种方式均受5个字段的上限限制——
```
add-field-suggestion.sh
```
会拒绝关联第6个字段。
Opportunity阶段：Opportunity阶段是选择列表值，不同的Opportunity记录类型可能显示不同的阶段子集。使用的是标准还是自定义Opportunity阶段？是否需要定义/更新它们的描述？注意：
```
OpptStageDescription
```
是每个阶段全局生效——仅由
```
OpportunityStageApiName
```
作为键，无按记录类型区分的列——因此每个活跃阶段的描述只需编写一次，并适用于所有显示该阶段的记录类型。无论组织是否有Opportunity记录类型，安装流程的操作方式相同。

Admin Communication Guidelines

管理员沟通指南

CRITICAL: This skill serves admin users, not developers. Follow these rules:

Run all bash commands in background (
```
run_in_background: true
```
) - this includes the main setup commands AND any log-tailing for progress monitoring
Monitor progress without showing commands - You can tail logs to provide progress updates, but run tail commands in background too. Parse the output and show only friendly updates like "✓ Opportunity Team: enabled". Never show the tail command itself or raw log lines.
Investigate errors silently - When something fails, do your diagnosis (check data, run queries, fix environment issues) in background without narrating each step. Only surface the conclusion and action: "I need to create a test opportunity first. Let me do that..." NOT "The --json output is corrupted... let me check... ANSI color codes... let me fix..."
Narrate in plain language: Say "Setting up the platform..." NOT "Running setup-all.sh Phase 1.5"
Hide technical details: No SOAP responses, curl commands, jq parsing, phase numbers, or script paths unless user asks to debug
Check exit codes: Always parse success/failure and translate technical errors to admin-friendly messages
Set time expectations: "This takes ~3 minutes..." prevents "is it frozen?" questions
Handle errors gracefully: Auto-retry transient failures; explain and offer choices for real blockers

See

references/admin-communication.md

for detailed error translation patterns, examples, and recovery strategies.

重要提示：本技能服务于管理员用户，而非开发人员。请遵循以下规则：

在后台运行所有bash命令（
```
run_in_background: true
```
）——包括主安装命令和任何用于监控进度的日志跟踪命令
监控进度但不显示命令——你可以跟踪日志以提供进度更新，但日志跟踪命令也需在后台运行。解析输出内容，仅显示友好的更新信息，如“✓ Opportunity Team：已启用”。永远不要显示跟踪命令本身或原始日志行。
静默排查错误——当出现故障时，在后台进行诊断（检查数据、运行查询、修复环境问题），不要叙述每一步操作。仅告知结论和操作：“我需要先创建一个测试Opportunity。我来处理……” 而不是“--json输出已损坏……我来检查……ANSI颜色代码……我来修复……”
用通俗易懂的语言叙述：说“正在设置平台……” 而不是“正在运行setup-all.sh第1.5阶段”
隐藏技术细节：除非用户要求调试，否则不要显示SOAP响应、curl命令、jq解析、阶段编号或脚本路径
检查退出码：始终解析成功/失败状态，并将技术错误转换为管理员友好的消息
设定时间预期：“这大约需要3分钟……” 可以避免用户询问“是不是卡住了？”
优雅处理错误：自动重试临时故障；对于真正的阻塞问题，说明情况并提供选择

详细的错误转换模式、示例和恢复策略请参见

references/admin-communication.md

。

Workflow

工作流程

IMPORTANT — Primary Entry Point: For end-to-end setup, drive

scripts/setup-all.sh

. It handles all prerequisites, enablement, flow deployment, permissions, and verification in the correct dependency order. Only use individual scripts when diagnosing or repairing specific components after

setup-all.sh

has run.

setup-all.sh

is a 3-phase, field-selection-driven setup: (1) platform enablement + agent user + PSG assignment; (2) field selection → prompt template creation/activation → stage descriptions (only if

StageName

selected) → prompt verification; (3) flow build (selected fields only), flow activation, agent activation, PSG recalc. It is safe by default: suggestions-only mode, no stage-description creation unless asked, and the flow does not go live until the prompts are in place.

重要提示——主要入口点：对于端到端设置，请运行

scripts/setup-all.sh

。它会按正确的依赖顺序处理所有前提条件、启用、Flow部署、权限和验证。仅当

setup-all.sh

运行后需要诊断或修复特定组件时，才使用单独的脚本。

setup-all.sh

是一个基于字段选择的3阶段安装流程：(1) 平台启用 + agent用户 + PSG分配；(2) 字段选择 → 提示模板创建/激活 → 阶段描述（仅当选择

StageName

时）→ 提示验证；(3) Flow构建（仅选中的字段）、Flow激活、agent激活、PSG重新计算。默认情况下是安全的：仅建议模式，除非明确要求否则不创建阶段描述，且Flow在提示模板就绪前不会生效。

Canonical agent path — two calls with an interactive tune-loop between them

标准agent流程——两次调用，中间包含交互式调优循环

See

references/canonical-agent-path.md

for the full two-call pattern with code examples, agent rules, narration templates, and internal step map.

Summary: Split setup into Call 1 (

--through-phase prompts

) and Call 2 (

--from-phase flow

) with a tune-loop between them where you drive prompt approval per field. Call 2 is MANDATORY — without it no flow exists and no suggestions generate.

完整的两次调用模式、代码示例、agent规则、叙述模板和内部步骤映射请参见

references/canonical-agent-path.md

。

摘要：将安装分为调用1（

--through-phase prompts

）和调用2（

--from-phase flow

），中间包含调优循环，在此循环中你可以按字段驱动提示审批。调用2是必须的——没有它，Flow不存在，也不会生成任何建议。

Unattended / CI path (single call, no tune-loop)

无人值守/CI流程（单次调用，无调优循环）

For fully unattended runs (cron/CI, or when the user explicitly declines prompt review):

bash

bash setup-all.sh <org-alias> --fields "NextStep,StageName,Risk__c" \
  [--field-goal "NextStep:<text>"] [--field-instruction "NextStep:<text>"] \
  [--autonomous] [--create-stage-descriptions] [--skip-prompt-verification] \
  --non-interactive [--users "a@x.com,b@x.com"]

In single-pass mode the verification step still prints each field's suggestion (informational). Relay them to the user.

对于完全无人值守的运行（cron/CI，或用户明确拒绝审核提示）：

bash

bash setup-all.sh <org-alias> --fields "NextStep,StageName,Risk__c" \\
  [--field-goal "NextStep:<text>"] [--field-instruction "NextStep:<text>"] \\
  [--autonomous] [--create-stage-descriptions] [--skip-prompt-verification] \\
  --non-interactive [--users "a@x.com,b@x.com"]

在单次运行模式下，验证步骤仍会打印每个字段的建议（仅供参考）。请将这些建议告知用户。

Flags

标志位

See

references/flags.md

for complete flag documentation. Key flags:

```
--check-license
```
— preflight only (exit 0 = capable, 1 = blocker). Run FIRST.
```
--through-phase prompts
```
/
```
--from-phase flow
```
— split-run endpoints (mutually exclusive)
```
--fields "NextStep,StageName"
```
— required field set (at least 1, max 5)
```
--autonomous
```
— enable auto-apply mode (off by default)
```
--non-interactive
```
/
```
--yes
```
— unattended mode (requires
```
--fields
```
)
```
--users "a@x.com,b@x.com"
```
— explicit user list for PSG assignment (email-validated)

The phases below explain the decision logic for manual/repair scenarios. See

references/setup-order.md

for step-by-step details.

完整的标志位文档请参见

references/flags.md

。关键标志位：

```
--check-license
```
— 仅预检（返回0表示具备条件，返回1表示存在阻塞）。请首先运行此命令。

--through-phase prompts

--from-phase flow

— 拆分运行的端点（互斥）

```
--fields "NextStep,StageName"
```
— 必填字段集（至少1个，最多5个）
```
--autonomous
```
— 启用自动应用模式（默认关闭）
```
--non-interactive
```
/
```
--yes
```
— 无人值守模式（需要
```
--fields
```
参数）
```
--users "a@x.com,b@x.com"
```
— 用于PSG分配的明确用户列表（需验证邮箱有效性）

以下阶段解释了手动/修复场景的决策逻辑。详细步骤请参见

references/setup-order.md

。

Automation Summary

自动化摘要

Every configuration step and whether it's CLI-automatable (and how) or UI-only lives in

references/automation-matrix.md

. In short: prerequisites, enablement, flow clone/activation, agent publish/activation, permissions, field suggestions, and stage descriptions are all CLI-automatable (

setup-all.sh

does them in dependency order); only Agent Analytics is UI-only.

每个配置步骤是否可通过CLI自动化（以及如何自动化）或仅可通过UI操作，请参见

references/automation-matrix.md

。简而言之：前提条件、启用、Flow克隆/激活、agent发布/激活、权限、字段建议和阶段描述均可通过CLI自动化（

setup-all.sh

按依赖顺序处理这些步骤）；仅Agent Analytics需通过UI操作。

Phase 0 — Authentication and Org Assessment

阶段0 — 身份验证和组织评估

Authenticate — See

references/auth-and-cli.md

. Always use

2>/dev/null

sf --json | jq

Assess org state — Branching signals (NOT PSG existence — PSGs ship with the license even in unconfigured orgs):
- Enablement:
```
SalesDealAgentSettings.enableDealAgent
```
  via SOAP readMetadata
- Setup-has-run (any of): (a) agent user holds
```
SalesManagementAgentUserPsg
```
  , (b) flow
```
Process_Field_Update_Suggestions
```
  exists (by ApiName, not label), (c)
```
BotDefinition
```
  with
```
AgentTemplate IN ('SalesMgmt__NGASalesAgent','SalesMgmt__SalesAgent')
```
Run
scripts/setup-all.sh <org-alias>
— handles branching automatically:
- No enablement + no signals → greenfield (Phase 1)
- Partial signals → repair (Phase 3)
- All present → verification only

身份验证 — 请参见
```
references/auth-and-cli.md
```
。在
```
sf --json | jq
```
命令中始终使用
```
2>/dev/null
```
。
评估组织状态 — 分支信号（不是PSG是否存在——即使在未配置的组织中，PSG也会随许可证一起提供）：
- 启用状态：通过SOAP readMetadata获取
```
SalesDealAgentSettings.enableDealAgent
```
- 已运行安装（满足以下任一条件）：(a) agent用户持有
```
SalesManagementAgentUserPsg
```
  ，(b) Flow
```
Process_Field_Update_Suggestions
```
  存在（通过ApiName判断，而非标签），(c)
```
BotDefinition
```
  的
```
AgentTemplate IN ('SalesMgmt__NGASalesAgent','SalesMgmt__SalesAgent')
```
运行
scripts/setup-all.sh <org-alias>
— 自动处理分支逻辑：
- 未启用 + 无信号 → 全新组织（阶段1）
- 部分信号 → 修复（阶段3）
- 所有信号都存在 → 仅验证

Phase 1 — Enable Prerequisites and Pipeline Management

阶段1 — 启用前提条件和Pipeline Management

See
references/setup-order.md
for complete scripts. Enable in dependency order: Einstein GenAI → Agentforce Agent → Agent Platform → Enhanced Notes → Opportunity Team → Pipeline Management → Pipeline Inspection. Then deploy the flow (

scripts/create-flow.sh

) and verify agent architecture (

references/agent-creation.md

What auto-creates on enablement: agent user + PSGs (auto-assigned). What you must create:

BotDefinition:SalesAgent

(via authoring-bundle publish) and the schedule-triggered flow (via template deploy).

完整脚本请参见
references/setup-order.md
。按依赖顺序启用：Einstein GenAI → Agentforce Agent → Agent Platform → Enhanced Notes → Opportunity Team → Pipeline Management → Pipeline Inspection。然后部署Flow（

scripts/create-flow.sh

）并验证agent架构（

references/agent-creation.md

）。

启用时自动创建的内容：agent用户 + PSG（自动分配）。必须创建的内容：

BotDefinition:SalesAgent

（通过创作包发布）和计划触发的Flow（通过模板部署）。

Phase 2 — Assign Permissions

阶段2 — 分配权限

Define Agent Access:
```
bash scripts/define-agent-access.sh <org-alias>
```
— creates custom permset, links into both PSGs. See
```
references/agent-creation.md
```
→ "Agent Access".
Assign PSG to target users: HARD RULE — never bulk-assign. Grant
```
SalesManagementUserPsg
```
only to: (1) the running user, and (2) explicit
```
--users
```
list. No enumerate-and-grant.

定义Agent访问权限：
```
bash scripts/define-agent-access.sh <org-alias>
```
— 创建自定义权限集，关联到两个PSG。请参见
```
references/agent-creation.md
```
→ "Agent Access"。
为目标用户分配PSG：严格规则——切勿批量分配。仅向以下用户授予
```
SalesManagementUserPsg
```
：(1) 当前运行用户，(2) 明确的
```
--users
```
列表。不要枚举所有用户并授予权限。

Phase 3 — Repair Mode (Partially Configured Org)

阶段3 — 修复模式（部分配置的组织）

See

references/repair-diagnostics.md

for the full checklist. Common issues: flow deactivated → Tooling API

PATCH

; agent not active →

sf agent activate

; prompt template inactive → version round-trip redeploy; agent missing → SOAP toggle off/on.

完整检查表请参见

references/repair-diagnostics.md

。常见问题：Flow已停用 → 使用Tooling API

PATCH

；agent未激活 →

sf agent activate

；提示模板未激活 → 版本往返重新部署；agent缺失 → SOAP开关关闭/开启。

Phase 4 — Customize (Admin Decisions Required)

阶段4 — 自定义（需要管理员决策）

Define Opportunity Stage Descriptions — Use the "propose and correct" pattern from
```
references/opportunity-stages.md
```
. Query active stages, present defaults from reference file, apply user corrections, bulk-create via Tooling API.
IMPORTANT: Stage descriptions MAY be auto-provisioned when Pipeline Management is enabled (test orgs showed MEDDIC descriptions pre-populated). Always CHECK for existing descriptions BEFORE attempting to create them. If descriptions already exist, UPDATE rather than CREATE to avoid duplicates.
Configure additional data sources — See
```
references/data-sources.md
```
for decision table (Einstein Conversation Insights, Einstein Activity Capture, Inbox).
Add custom opportunity field suggestions (AFTER setup) — Fields chosen at setup time are wired by
```
setup-all.sh --fields
```
. To add a new field to an already-configured org, run
```
scripts/add-field-suggestion.sh <org-alias> <FieldApiName>
```
(e.g. a
```
Risk__c
```
field). The script validates the field, fills the canonical template (
```
assets/field-completion-template.genAiPromptTemplate-meta.xml
```
), deploys and activates it via the version round-trip, and idempotently wires the field into the live
```
Process_Field_Update_Suggestions
```
flow (respecting the 5-field cap). Optional flags:
```
--label
```
,
```
--goal
```
,
```
--instruction
```
(the two field-specific prompt lines),
```
--verify-with-note
```
(seed a Note + run a synchronous generation),
```
--opp <Id>
```
,
```
--skip-flow
```
,
```
--force
```
. See
```
references/field-completion-prompts.md
```
for mechanics and
```
references/field-completion-prompt-template.md
```
for writing the goal/instruction lines.
Naming note: the prompt-template
```
<type>
```
is
```
einstein_gpt__fieldCompletion
```
(used in all metadata — never change it). The Setup UI (Prompt Builder) labels this category "Field Generation" — same thing. Admins should look for "Field Generation", not "Field Completion".
Configure autonomous updates — Autonomous mode is a single org-wide toggle (
```
SalesDealAgentSettings.enableDealAgentAutoApproveAllTasks
```
), not a per-field setting: when ON, the agent auto-applies all field suggestions without review; when OFF (the default), every suggestion waits for human approval. It is opt-in via
```
setup-all.sh --autonomous
```
. See
```
references/autonomous-updates.md
```
for the toggle and safety considerations.
Enable Agent Analytics — Setup → Einstein Feedback and Monitoring → Agent Analytics (UI-only).

定义Opportunity阶段描述 — 使用
```
references/opportunity-stages.md
```
中的“提议并修正”模式。查询活跃阶段，提供参考文件中的默认描述，应用用户的修正，通过Tooling API批量创建。
重要提示：启用Pipeline Management时，阶段描述可能会自动预配置（测试组织显示MEDDIC描述已预先填充）。在尝试创建描述之前，务必检查是否已有描述。如果描述已存在，请更新而非创建，以避免重复。
配置额外数据源 — 决策表请参见
```
references/data-sources.md
```
（Einstein Conversation Insights、Einstein Activity Capture、Inbox）。
添加自定义Opportunity字段建议（安装完成后） — 安装时选择的字段会由
```
setup-all.sh --fields
```
关联。要向已配置的组织添加新字段，请运行
```
scripts/add-field-suggestion.sh <org-alias> <FieldApiName>
```
（例如
```
Risk__c
```
字段）。脚本会验证字段，填充标准模板（
```
assets/field-completion-template.genAiPromptTemplate-meta.xml
```
），通过版本往返部署并激活，然后幂等性地将字段关联到已启用的
```
Process_Field_Update_Suggestions
```
Flow（遵守5个字段的上限）。可选标志位：
```
--label
```
、
```
--goal
```
、
```
--instruction
```
（两个字段特定的提示行）、
```
--verify-with-note
```
（添加一个Note并运行同步生成）、
```
--opp <Id>
```
、
```
--skip-flow
```
、
```
--force
```
。机制请参见
```
references/field-completion-prompts.md
```
，编写目标/指令行请参见
```
references/field-completion-prompt-template.md
```
。
命名说明：提示模板的
```
<type>
```
为
```
einstein_gpt__fieldCompletion
```
（所有元数据中均使用此值——切勿修改）。设置UI（Prompt Builder）将此类别标记为**"Field Generation"**——两者含义相同。管理员应查找"Field Generation"，而非"Field Completion"。
配置自动更新 — 自动模式是一个全组织范围的开关（
```
SalesDealAgentSettings.enableDealAgentAutoApproveAllTasks
```
），而非按字段设置：开启时，agent会自动应用所有字段建议，无需审核；关闭时（默认），所有建议需等待人工批准。可通过
```
setup-all.sh --autonomous
```
选择启用。开关和安全注意事项请参见
```
references/autonomous-updates.md
```
。
启用Agent Analytics — 设置 → Einstein Feedback and Monitoring → Agent Analytics（仅UI操作）。

Understanding the Flows

理解Flow

See

references/flow-clone-from-template.md

for complete flow architecture, clone approach, and technical constraints.

Key facts: Pipeline Management uses a schedule-triggered flow (

Process_Field_Update_Suggestions

) cloned from the managed template. Detection is by ApiName (not label). The flow calls

getOrExecFieldUpdtSuggestion

with the selected fields — adding a field requires both activating its template AND wiring it into the flow's collection (handled atomically by

add-field-suggestion.sh

). No suggestions generate until this flow is active.

完整的Flow架构、克隆方法和技术约束请参见

references/flow-clone-from-template.md

。

关键事实：Pipeline Management使用从托管模板克隆的计划触发Flow（

Process_Field_Update_Suggestions

）。通过ApiName检测（而非标签）。Flow调用

getOrExecFieldUpdtSuggestion

并传入选中的字段——添加字段需要激活其模板并将其关联到Flow的集合（由

add-field-suggestion.sh

原子性处理）。只有当此Flow激活后，才会生成建议。

Key Considerations

关键注意事项

Pipeline Management REQUIRES a standalone
BotDefinition:SalesAgent
— a missing BotDefinition blocks setup. Auto-provisioning creates it in most editions, but where it doesn't,
```
publish_and_activate_agent()
```
(in
```
shared/agent-bundle-publish.sh
```
, called by
```
setup-all.sh
```
and
```
create-agent.sh
```
) publishes the bundle from
```
assets/sales_management_agent.agent
```
and activates it. Do NOT rely on the
```
EmployeeCopilotPlanner
```
fallback — users need the interactive chat surface.
Flow runs only when agent is active with Read/Write on opportunities; agent joins as opportunity team member (watch for Apex triggers on OpportunityTeamMember)
Daily limit: 8,000 LLM requests / ~4,000 opportunities. Suggestions expire after 30 days
Stage template reads
```
OpptStageDescription
```
via
```
GetOpportunityStageDetailsInvocableAction
```
— stage suggestions fail if any active stage lacks a description
Do NOT use CLI Metadata deploy for DealAgent enablement — SOAP API v64.0 only (CLI has silent failure mode)
Do NOT enable
BotSettings
/
enableBots
— legacy messaging bots, unrelated, fails with "Legal Terms" error
Pipeline Management does NOT use a managed package — components provisioned directly on enablement
Toggle not visible in Setup → verify Sales Cloud EE+ license and Agentforce SKU
Always use
```
2>/dev/null
```
on
```
sf ... --json
```
piped to jq (strips CLI warnings that corrupt JSON)

Pipeline Management需要独立的
BotDefinition:SalesAgent
— 缺失BotDefinition会阻止安装。大多数版本中自动资源调配会创建它，但如果未创建，
```
publish_and_activate_agent()
```
（位于
```
shared/agent-bundle-publish.sh
```
，由
```
setup-all.sh
```
和
```
create-agent.sh
```
调用）会从
```
assets/sales_management_agent.agent
```
发布包并激活。不要依赖
EmployeeCopilotPlanner
作为备选——用户需要交互式聊天界面。
Flow仅在agent激活且对Opportunity拥有读写权限时运行；agent会加入Opportunity团队（注意OpportunityTeamMember上的Apex触发器）
每日限制：8,000次LLM请求 / ~4,000个Opportunity。建议30天后过期
阶段模板通过
```
GetOpportunityStageDetailsInvocableAction
```
读取
```
OpptStageDescription
```
— 如果任何活跃阶段缺少描述，阶段建议会失败
请勿使用CLI Metadata部署来启用DealAgent — 仅支持SOAP API v64.0（CLI存在静默失败模式）
请勿启用
BotSettings
/
enableBots
— 这是旧版消息bot，与本功能无关，会触发“法律条款”错误
Pipeline Management不使用托管包 — 组件在启用时直接预配置
设置中看不到开关 → 验证Sales Cloud EE+许可证和Agentforce SKU
在
```
sf ... --json
```
管道到jq的命令中始终使用
```
2>/dev/null
```
（去除会破坏JSON的CLI警告）

References

参考资料

```
references/setup-order.md
```
— Complete copy-paste setup sequence with all scripts
```
references/soap-api-enablement.md
```
— SOAP API patterns for all prerequisites
```
references/flow-clone-from-template.md
```
— Flow clone approach with ApiName-based detection
```
references/agent-creation.md
```
— Agent creation (auto-provisioning + authoring-bundle publish; the SOAP toggle only re-provisions the agent user and PSGs, not the BotDefinition)
```
references/field-completion-prompts.md
```
— Adding custom field suggestions end-to-end (verified v67 pattern, the three requirements,
```
add-field-suggestion.sh
```
mechanics, verification)
```
references/field-completion-prompt-template.md
```
— Writing the field-specific goal/instruction lines the script consumes
```
references/opportunity-stages.md
```
— Stage descriptions + methodologies
```
references/repair-diagnostics.md
```
— Partially configured org handling
```
references/auth-and-cli.md
```
— Auth methods, CLI compatibility
```
references/data-sources.md
```
— Decision table for EAC/ECI/Notes/Enhanced Email
```
references/autonomous-updates.md
```
— Auto-apply mode toggle and safety considerations
```
references/metadata-inventory.md
```
— Complete inventory of provisioned metadata components
```
references/automation-matrix.md
```
— Which setup steps are CLI-automatable (and how) vs UI-only
```
references/admin-communication.md
```
— Admin-friendly narration patterns: error-translation table, silent-investigation examples, and per-step narration templates for driving setup conversationally

```
references/setup-order.md
```
— 完整的可复制粘贴安装序列及所有脚本
```
references/soap-api-enablement.md
```
— 所有前提条件的SOAP API模式
```
references/flow-clone-from-template.md
```
— 基于ApiName检测的Flow克隆方法
```
references/agent-creation.md
```
— Agent创建（自动资源调配 + 创作包发布；SOAP开关仅重新预配置agent用户和PSG，不包括BotDefinition）
```
references/field-completion-prompts.md
```
— 端到端添加自定义字段建议（已验证的v67模式、三个要求、
```
add-field-suggestion.sh
```
机制、验证）
```
references/field-completion-prompt-template.md
```
— 编写脚本所需的字段特定目标/指令行
```
references/opportunity-stages.md
```
— 阶段描述 + 方法学
```
references/repair-diagnostics.md
```
— 部分配置组织的处理
```
references/auth-and-cli.md
```
— 身份验证方法、CLI兼容性
```
references/data-sources.md
```
— EAC/ECI/Notes/Enhanced Email的决策表
```
references/autonomous-updates.md
```
— 自动应用模式开关和安全注意事项
```
references/metadata-inventory.md
```
— 预配置元数据组件的完整清单
```
references/automation-matrix.md
```
— 哪些安装步骤可通过CLI自动化（以及如何自动化） vs 仅可通过UI操作
```
references/admin-communication.md
```
— 管理员友好的叙述模式：错误转换表、静默排查示例、以及用于引导安装对话的逐步骤叙述模板

Examples

示例

```
examples/custom-prompt-instructions.md
```
— Custom prompt instruction patterns for methodology alignment and update aggressiveness control

```
examples/custom-prompt-instructions.md
```
— 用于方法学对齐和更新力度控制的自定义提示指令模式

Standalone Scripts

独立脚本

Executable scripts for each setup phase (run from

scripts/

directory with

sfdx-project.json

in CWD):

```
scripts/setup-all.sh
```
— Master orchestration: 3-phase, field-selection-driven end-to-end setup (auth + enablement + agent user/PSG → field selection + prompt templates + stage descriptions + prompt-verification gate → flow build/activate + agent activate + PSG recalc). Requires
```
--fields
```
(aborts if none); interactive by default; safe defaults (suggestions-only, no stage-description creation) with
```
--field-goal
```
/
```
--field-instruction
```
,
```
--autonomous
```
,
```
--create-stage-descriptions
```
,
```
--skip-prompt-verification
```
opt-in flags and
```
--non-interactive
```
/
```
--yes
```
for unattended runs
```
scripts/shared/flow-builder.sh
```
— flow field-collection helpers sourced by
```
setup-all.sh
```
:
```
build_field_collection
```
(greenfield — writes a flow wiring only the selected fields), plus
```
strip_flow_field
```
/
```
add_flow_field
```
/
```
flow_wired_fields
```
(repair — reconcile an already-deployed flow's fields in place)

scripts/shared/stage-descriptions.sh

— Phase 4c.5 stage-description creation sourced by

setup-all.sh

; exports

run_stage_descriptions()

(self-gates on

STAGENAME_SELECTED

; sets

STAGE_DESCRIPTIONS_BLOCKED

for the Phase 8.5 summary)

```
scripts/shared/test-opp.sh
```
— shared test-opportunity + grounding-note resolution sourced by both verify scripts (
```
flow-debug-and-verify.sh
```
,
```
verify-prompt-generation.sh
```
); exports
```
resolve_test_opportunity
```
(reuse/create/fallback the
```
[PM-TEST]
```
test opp, roll CloseDate, seed/refresh the grounding note) and
```
seed_grounding_note
```
```
scripts/retrieve-settings.sh
```
— Audit current org settings state
```
scripts/enable-prerequisites.sh
```
— Enable prerequisite settings only (Einstein GenAI, Agentforce, Enhanced Notes, Opportunity Team) via SOAP — no fields/flow/agent.
```
setup-all.sh
```
runs this same enablement inline; use the standalone script only for surgical, non-destructive prerequisite toggling
```
scripts/enable-deal-agent.sh
```
— Enable Pipeline Management via SOAP API (standalone helper;
```
setup-all.sh
```
enables PM inline — use this only for isolated manual toggling)
```
scripts/create-flow.sh
```
— Detect, deploy from template, or activate suggestion flow (falls back to UI guidance)
```
scripts/create-agent.sh
```
— Verify/create/activate agent (auto-provisioning + authoring-bundle publish; SOAP toggle only re-provisions agent user/PSGs, not the BotDefinition); also defines Agent Access on completion
```
scripts/define-agent-access.sh
```
— Define Agent Access so both users holding
```
SalesManagementUserPsg
```
and the autonomous agent user holding
```
SalesManagementAgentUserPsg
```
can launch/run the agent (custom permset +
```
SetupEntityAccess
```
+
```
PermissionSetGroupComponent
```
into both PSGs + recalc)
```
scripts/deploy-settings.sh
```
— Deploy non-SOAP settings (Notes, Email, Opportunity)
```
scripts/add-field-suggestion.sh
```
— Add an AI field-completion suggestion for any eligible Opportunity text field (validate → fill canonical template → deploy + activate via version round-trip → wire flow;
```
--verify-with-note
```
for synchronous generation)
```
scripts/verify-all.sh
```
— Comprehensive configuration status check (read-only)

每个安装阶段的可执行脚本（从

scripts/

sfdx-project.json

）：

```
scripts/setup-all.sh
```
— 主编排脚本：基于字段选择的3阶段端到端安装（身份验证 + 启用 + agent用户/PSG → 字段选择 + 提示模板 + 阶段描述 + 提示验证门控 → Flow构建/激活 + agent激活 + PSG重新计算）。需要
```
--fields
```
参数（未提供则中止）；默认交互式；安全默认值（仅建议模式，不创建阶段描述），支持
```
--field-goal
```
/
```
--field-instruction
```
、
```
--autonomous
```
、
```
--create-stage-descriptions
```
、
```
--skip-prompt-verification
```
可选标志位，以及
```
--non-interactive
```
/
```
--yes
```
用于无人值守运行
```
scripts/shared/flow-builder.sh
```
—
```
setup-all.sh
```
引用的Flow字段集合助手：
```
build_field_collection
```
（全新组织——仅为选中的字段编写Flow关联），以及
```
strip_flow_field
```
/
```
add_flow_field
```
/
```
flow_wired_fields
```
（修复——协调已部署Flow的字段）

scripts/shared/stage-descriptions.sh

— 第4c.5阶段的阶段描述创建脚本，由

setup-all.sh

引用；导出

run_stage_descriptions()

（仅当

STAGENAME_SELECTED

时运行；为第8.5阶段摘要设置

STAGE_DESCRIPTIONS_BLOCKED

）

```
scripts/shared/test-opp.sh
```
— 两个验证脚本（
```
flow-debug-and-verify.sh
```
、
```
verify-prompt-generation.sh
```
）引用的共享测试Opportunity + 基础Note解析脚本；导出
```
resolve_test_opportunity
```
（重用/创建/回退
```
[PM-TEST]
```
测试Opportunity，将CloseDate延后30天，添加/刷新基础Note）和
```
seed_grounding_note
```
```
scripts/retrieve-settings.sh
```
— 审核当前组织的设置状态
```
scripts/enable-prerequisites.sh
```
— 仅启用前提条件设置（Einstein GenAI、Agentforce、Enhanced Notes、Opportunity Team），通过SOAP API——不处理字段/Flow/agent。
```
setup-all.sh
```
会在内部运行相同的启用流程；仅在需要精准、非破坏性的前提条件开关时使用此独立脚本
```
scripts/enable-deal-agent.sh
```
— 通过SOAP API启用Pipeline Management（独立助手脚本；
```
setup-all.sh
```
会在内部启用PM——仅在需要单独手动开关时使用）
```
scripts/create-flow.sh
```
— 检测、从模板部署或激活建议Flow（回退到UI指导）
```
scripts/create-agent.sh
```
— 验证/创建/激活agent（自动资源调配 + 创作包发布；SOAP开关仅重新预配置agent用户/PSG，不包括BotDefinition）；完成后还会定义Agent访问权限
```
scripts/define-agent-access.sh
```
— 定义Agent访问权限，使持有
```
SalesManagementUserPsg
```
的用户和持有
```
SalesManagementAgentUserPsg
```
的自动agent用户都能启动/运行agent（自定义权限集 +
```
SetupEntityAccess
```
+
```
PermissionSetGroupComponent
```
关联到两个PSG + 重新计算）
```
scripts/deploy-settings.sh
```
— 部署非SOAP设置（Notes、Email、Opportunity）
```
scripts/add-field-suggestion.sh
```
— 为任何符合条件的Opportunity文本字段添加AI字段补全建议（验证 → 填充标准模板 → 通过版本往返部署 + 激活 → 关联Flow；
```
--verify-with-note
```
用于同步生成）
```
scripts/verify-all.sh
```
— 全面的配置状态检查（只读）

Admin Verification Tools

管理员验证工具

The following scripts help admins verify that Pipeline Management is working correctly after setup. Use these to confirm field suggestions are generating as expected.

Before running either verify script, compose a realistic grounding note and pass it via the
PM_GROUNDING_NOTE
env var. The AI only generates a suggestion when the opportunity's Note carries real sales context — a generic placeholder grounds on nothing and returns empty, making a working setup look broken. Write a few sentences of a plausible sales call/meeting reflecting the configured

--fields

and any

--field-goal

--field-instruction

(e.g. concrete next steps and deadlines for

NextStep

; buyer, budget, pain, and timeline signals for

StageName

). Do not use a rigid template — tailor it to what was configured. Example:

PM_GROUNDING_NOTE="Discovery call with the VP of Sales: confirmed a $250K budget and a Q3 go-live; agreed next steps are a technical deep-dive with IT and an ROI analysis." bash scripts/flow-debug-and-verify.sh <org>

. Absent the env var, the scripts fall back to a generic note.

Both scripts are invoked separately by the skill but converge on the SAME test opportunity and the SAME grounding note. They both find-or-create a user-owned opp named

[PM-TEST] Test Pipeline Verification Opp

(the

[PM-TEST]

prefix makes every opp this tooling creates findable via

Opportunity WHERE Name LIKE '[PM-TEST]%'

). On reuse they roll the opp's CloseDate +30 days so it stays flow-eligible, and refresh the grounding note only when the

PM_GROUNDING_NOTE

you pass differs from the stored one (otherwise the existing note is left untouched). Pass the same

PM_GROUNDING_NOTE

value to both invocations so the note is consistent regardless of run order.

```
scripts/flow-debug-and-verify.sh
```
— Guided two-phase verification of the scheduled suggestion flow.
- flow-debug-start (
```
bash scripts/flow-debug-and-verify.sh <org> [opp-id]
```
  ) reuses a stable user-owned test opportunity named
```
[PM-TEST] Test Pipeline Verification Opp
```
  if one is still eligible (so repeated runs don't pile up duplicate opps), otherwise creates it, falls back to any user-owned open opp if creation fails, seeds/refreshes a grounding Note on whichever opp resolves, and prints the flow's Debug URL, the opportunity to select, and the exact
```
verify
```
  command to run afterward — including a trailing baseline count (the pre-run suggestion count). It uses a user-owned opp so the resulting suggestion is visible in the admin's owner-scoped Pipeline Inspection. If no opportunity is available it prints an actionable message and exits 0 without verifying. Relay the URL + opportunity to the user and wait for them to run Debug in Flow Builder against that opportunity — this is the only on-demand path that binds a real
```
$Record
```
  on a scheduled record-triggered flow (leave rollback mode UNCHECKED so the DML persists).
- verify (
```
bash scripts/flow-debug-and-verify.sh <org> <opp-id> verify [baseline]
```
  ) runs after the user confirms the debug run finished: it confirms the flow reached its action (agent user added to the Opportunity Team) and polls (~4 min) for the generated
```
AiGenActionItem
```
  suggestions. Pass the baseline that flow-debug-start printed — it is captured before the debug run, so the delta is exact; omitting it falls back to weaker absolute detection. The wait between phases happens in the conversation, not inside the script.
```
scripts/verify-prompt-generation.sh
```
— Test the prompt template behind a field by calling the Einstein
```
/generations
```
API directly. Pass the field API name (
```
bash scripts/verify-prompt-generation.sh <org> <field> [opp-id]
```
— e.g.
```
NextStep
```
,
```
Risk__c
```
); it resolves the field to its managed/derived template (a bare, un-namespaced managed name returns
```
ENTITY_IS_DELETED
```
) and uses the same user-owned opp selection strategy as
```
flow-debug-and-verify.sh
```
(both share one sourced lib: reuse the shared
```
[PM-TEST] Test Pipeline Verification Opp
```
→ create it → owner-scoped fallback → graceful skip) — the only difference on skip is that no Einstein endpoint call is made. Shows the actual suggestion text generated, helping admins validate custom prompt instructions.

以下脚本可帮助管理员在安装完成后验证Pipeline Management是否正常工作。使用这些脚本确认字段建议是否按预期生成。

运行任一验证脚本前，请编写一个真实的基础Note，并通过
PM_GROUNDING_NOTE
环境变量传递。只有当Opportunity的Note包含真实销售上下文时，AI才会生成建议——通用占位符无法提供有效基础，会导致返回空内容，使正常的安装看起来像故障。编写几句合理的销售通话/会议内容，反映已配置的

--fields

和任何

--field-goal

--field-instruction

（例如

NextStep

的具体后续步骤和截止日期；

StageName

的买方、预算、痛点和时间线信号）。不要使用僵化的模板——根据配置内容调整。示例：

PM_GROUNDING_NOTE="与销售副总裁的发现通话：确认25万美元预算和第三季度上线；同意后续步骤是与IT进行技术深度交流并进行ROI分析。" bash scripts/flow-debug-and-verify.sh <org>

。如果未提供环境变量，脚本会回退到通用Note。

两个脚本由技能单独调用，但会使用相同的测试Opportunity和相同的基础Note。 它们都会查找或创建名为

[PM-TEST] Test Pipeline Verification Opp

的用户所有Opportunity（

[PM-TEST]

前缀使此工具创建的所有Opportunity都可通过

Opportunity WHERE Name LIKE '[PM-TEST]%'

找到）。重用时会将Opportunity的CloseDate延后30天，使其保持Flow eligibility，仅当你传递的

PM_GROUNDING_NOTE

与存储的Note不同时才刷新基础Note（否则保留现有Note）。为两次调用传递相同的

PM_GROUNDING_NOTE

值，确保无论运行顺序如何，Note内容一致。

```
scripts/flow-debug-and-verify.sh
```
— 计划建议Flow的引导式两阶段验证。
- flow-debug-start（
```
bash scripts/flow-debug-and-verify.sh <org> [opp-id]
```
  ）会重用仍符合条件的稳定用户所有测试Opportunity（
```
[PM-TEST] Test Pipeline Verification Opp
```
  ），避免重复运行产生重复Opportunity；否则创建它；如果创建失败，回退到任何用户所有的未关闭Opportunity；在解析到的Opportunity上添加/刷新基础Note；然后打印Flow的调试URL、要选择的Opportunity，以及后续要运行的精确
```
verify
```
  命令——包括末尾的基线计数（运行前的建议计数）。它使用用户所有的Opportunity，使生成的建议在管理员的所有者范围Pipeline Inspection中可见。如果没有可用的Opportunity，会打印可操作的消息并返回0，不进行验证。将URL + Opportunity告知用户，等待他们在Flow Builder中针对该Opportunity运行调试——这是唯一能在计划记录触发Flow上绑定真实
```
$Record
```
  的按需路径（不要选中回滚模式，以便DML操作持久化）。
- verify（
```
bash scripts/flow-debug-and-verify.sh <org> <opp-id> verify [baseline]
```
  ）在用户确认调试运行完成后运行：确认Flow已执行到其操作（agent用户已添加到Opportunity Team），并轮询（约4分钟）生成的
```
AiGenActionItem
```
  建议。传递flow-debug-start打印的基线计数——它是在调试运行前捕获的，因此差值准确；如果省略，会回退到较弱的绝对检测。两个阶段之间的等待在对话中进行，而非脚本内部。
```
scripts/verify-prompt-generation.sh
```
— 通过直接调用Einstein
```
/generations
```
API测试字段背后的提示模板。传递字段API名称（
```
bash scripts/verify-prompt-generation.sh <org> <field> [opp-id]
```
——例如
```
NextStep
```
、
```
Risk__c
```
）；它会将字段解析为其托管/派生模板（未命名空间的托管名称会返回
```
ENTITY_IS_DELETED
```
），并使用与
```
flow-debug-and-verify.sh
```
相同的用户所有Opportunity选择策略（两者共享一个引用库：重用共享的
```
[PM-TEST] Test Pipeline Verification Opp
```
→ 创建它 → 所有者范围回退 → 优雅跳过）——唯一的区别是跳过不会调用Einstein端点。显示生成的实际建议文本，帮助管理员验证自定义提示指令。

Agent Assets

Agent资源

```
assets/sales_management_agent.agent
```
— Agent Script definition for the Sales Management Agent (used when auto-provisioning fails)

assets/sales_management_agent.bundle-meta.xml

— AiAuthoringBundle metadata (deploy with

sf agent publish authoring-bundle --api-name SalesAgent

)

assets/field-completion-template.genAiPromptTemplate-meta.xml

— Canonical

einstein_gpt__fieldCompletion

prompt template with

@@PLACEHOLDERS@@

, filled in by

scripts/add-field-suggestion.sh

(do not deploy directly)

```
assets/sales_management_agent.agent
```
— Sales Management Agent的Agent脚本定义（自动资源调配失败时使用）

assets/sales_management_agent.bundle-meta.xml

— AiAuthoringBundle元数据（使用

sf agent publish authoring-bundle --api-name SalesAgent

部署）

assets/field-completion-template.genAiPromptTemplate-meta.xml

— 标准

einstein_gpt__fieldCompletion

提示模板，带有

@@PLACEHOLDERS@@

，由

scripts/add-field-suggestion.sh

填充（请勿直接部署）",

sales-agentforce-pipeline-management-configure

Original

Translation

Agentforce Pipeline Management Configuration

Agentforce Pipeline Management 配置

Scope

适用范围

Prerequisites

前提条件

You must verify (scripts cannot check or change these)

必须验证（脚本无法检查或修改这些内容）

Enabled automatically by scripts/setup-all.sh

由scripts/setup-all.sh自动启用

Clarifying Questions

澄清问题

Admin Communication Guidelines

管理员沟通指南

Workflow

工作流程

Canonical agent path — two calls with an interactive tune-loop between them

标准agent流程——两次调用，中间包含交互式调优循环

Unattended / CI path (single call, no tune-loop)

无人值守/CI流程（单次调用，无调优循环）

Flags

标志位

Automation Summary

自动化摘要

Phase 0 — Authentication and Org Assessment

阶段0 — 身份验证和组织评估

Phase 1 — Enable Prerequisites and Pipeline Management

阶段1 — 启用前提条件和Pipeline Management

Phase 2 — Assign Permissions

阶段2 — 分配权限

Phase 3 — Repair Mode (Partially Configured Org)

阶段3 — 修复模式（部分配置的组织）

Phase 4 — Customize (Admin Decisions Required)

阶段4 — 自定义（需要管理员决策）

Understanding the Flows

理解Flow

Key Considerations

关键注意事项

References

参考资料

Examples

示例

Standalone Scripts

独立脚本

Admin Verification Tools

管理员验证工具

Agent Assets

Agent资源

Enabled automatically by
`scripts/setup-all.sh`

由
`scripts/setup-all.sh`
自动启用