Token导航 LogoToken导航

Agent Skill · 研究检索

sf-omniscript(科幻全向文字)

sf-omniscript 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

sf-omniscript 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为需要联网;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 62,001。

总安装 343周安装 14GitHub Stars 401下载 111核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:sf-omniscript(科幻全向文字)
来源仓库:https://github.com/jaganpro/sf-skills
仓库路径:skills/sf-omniscript
安装命令:
npx skills add https://github.com/jaganpro/sf-skills --skill sf-omniscript
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/jaganpro/sf-skills --skill sf-omniscript

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

用于查找、检索和筛选相关信息。sf-omniscript 属于研究检索类 Skill,可作为该场景下的辅助能力补充。适合根据关键词或任务场景快速定位候选结果。可结合来源仓库和原始 README 核验具体用法。安装前建议确认权限范围、维护状态及是否触发联网或命令执行。

安装方式:通过 npx 从指定 GitHub 仓库添加技能。

来源仓库内路径:skills/sf-omniscript

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/jaganpro/sf-skills --skill sf-omniscript;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要需要联网,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

sf-omniscript 怎么安装?

在终端执行 npx skills add https://github.com/jaganpro/sf-skills --skill sf-omniscript。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

sf-omniscript 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

sf-omniscript 需要什么权限?

权限级别是需要联网。该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

sf-omniscript 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

sf-omniscript 用的人多吗?

公开数据是累计安装 343 次,近一周安装 14 次,下载 111 次,GitHub 401 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

sf-omniscript 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
sf-omniscript

科幻全向文字 · 本页

343401需要联网未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

sf-omniscript: OmniStudio OmniScript Creation and Validation

Expert OmniStudio OmniScript builder for declarative, step-based guided digital experiences. OmniScripts are the OmniStudio analog of Screen Flows: multi-step, interactive processes that collect input, orchestrate server-side logic (Integration Procedures, DataRaptors), and present results to the user — all without code.

Quick Reference

Scoring: 120 points across 6 categories. Thresholds: ✅ 90+ (Deploy) | ⚠️ 67-89 (Review) | ❌ <67 (Block - fix required)


Core Responsibilities

  1. OmniScript Generation: Create well-structured OmniScripts from requirements, selecting appropriate element types for each step
  2. Element Design: Configure PropertySetConfig JSON for each element with correct data binding, validation, and conditional logic
  3. Dependency Analysis: Map all references to Integration Procedures, DataRaptors, and embedded OmniScripts before deployment
  4. Data Flow Analysis: Trace data through the OmniScript JSON structure — from prefill through user input to final save actions

CRITICAL: Orchestration Order

sf-omnistudio-analyze → sf-datamapper → sf-integration-procedure → sf-omniscript → sf-flexcard (you are here: sf-omniscript)

OmniScripts consume Integration Procedures and DataRaptors. Build those FIRST. FlexCards may launch OmniScripts — build FlexCards AFTER. Use sf-omnistudio-analyze to map the full dependency tree before starting.


Key Insights

InsightDetails
Type/SubType/Language tripletUniquely identifies an OmniScript. All three values are required and form the composite key. Example: Type=ServiceRequest, SubType=NewCase, Language=English
PropertySetConfigJSON blob containing all element configuration — layout, data binding, validation rules, conditional visibility. This is where the real logic lives
Core namespaceOmniProcess with IsIntegrationProcedure = false (equivalently OmniProcessType='OmniScript'). Elements are child OmniProcessElement records
Element hierarchyElements use Level/Order fields for tree structure. Level 0 = Steps, Level 1+ = elements within steps. Order determines sequence within a level
Version managementMultiple versions can exist; only one can be active per Type/SubType/Language triplet. Activate via the IsActive field
Data JSONOmniScripts pass a single JSON data structure through all steps. Elements read from and write to this shared JSON via merge field syntax

Workflow Design (5-Phase Pattern)

Phase 1: Requirements Gathering

Before building, evaluate alternatives: OmniScripts are best for complex, multi-step guided processes. For simple single-screen data entry, consider Screen Flows. For data display without interaction, consider FlexCards.

Ask the user to gather:

  • Type: The process category (e.g., ServiceRequest, Enrollment, ClaimSubmission)
  • SubType: The specific variation (e.g., NewCase, UpdateAddress, FileAppeal)
  • Language: Typically English unless multi-language support is required
  • Purpose: What business process this OmniScript guides the user through
  • Target org: Org alias for deployment
  • Data sources: Which objects/APIs need to be queried or updated

Then: Check existing OmniScripts to avoid duplication, identify reusable Integration Procedures or DataRaptors, and map the dependency chain.

Phase 2: Design & Element Selection

Design each step and select element types appropriate to the interaction pattern.

Container Elements
Element TypePurposeKey Config
StepTop-level container for a group of UI elements; each Step is a page in the wizardchartLabel, knowledgeOptions, show (conditional visibility)
Conditional BlockShow/hide a group of elements based on conditionsconditionType, show expression
Loop BlockIterate over a data list and render elements for each itemloopData (JSON path to array)
Edit BlockInline editing container for tabular dataeditFields, dataSource
Input Elements
Element TypePurposeKey Config
TextSingle-line text inputlabel, placeholder, pattern (regex validation)
Text AreaMulti-line text inputlabel, maxLength, rows
NumberNumeric input with optional formattinglabel, min, max, step, format
DateDate pickerlabel, dateFormat, minDate, maxDate
Date/TimeDate and time pickerlabel, dateFormat, timeFormat
CheckboxBoolean togglelabel, defaultValue
RadioRadio button group for single selectionlabel, options (static or data-driven)
SelectDropdown selectionlabel, options, optionSource (static/data)
Multi-selectMultiple item selectionlabel, options, maxSelections
Type AheadSearch/autocomplete inputlabel, dataSource, searchField, minCharacters
SignatureSignature capture padlabel, penColor, backgroundColor
FileFile uploadlabel, maxFileSize, allowedExtensions
CurrencyCurrency input with locale formattinglabel, currencyCode, min, max
EmailEmail input with format validationlabel, placeholder
TelephonePhone number input with maskinglabel, mask, placeholder
URLURL input with format validationlabel, placeholder
PasswordMasked text inputlabel, minLength
RangeSlider inputlabel, min, max, step
TimeTime pickerlabel, timeFormat
Display Elements
Element TypePurposeKey Config
Text BlockStatic content display (HTML supported)textContent, HTMLTemplateId
HeadlineSection headingtext, level (h1-h6)
AggregateCalculated summary displayaggregateExpression, format
DisclosureExpandable/collapsible contentlabel, defaultExpanded
ImageImage displayimageURL, altText
ChartData visualizationchartType, dataSource
Action Elements
Element TypePurposeKey Config
DataRaptor Extract ActionPull data from Salesforcebundle, inputMap, outputMap
DataRaptor Load ActionPush data to Salesforcebundle, inputMap
Integration Procedure ActionCall server-side Integration ProcedureipMethod (Type_SubType), inputMap, outputMap, remoteOptions
Remote ActionCall Apex @RemoteAction or RESTremoteClass, remoteMethod, inputMap
Navigate ActionPage navigation or redirectiontargetType, targetId, URL
DocuSign Envelope ActionTrigger DocuSign envelopetemplateId, recipientMap
Email ActionSend emailemailTemplateId, recipientMap
Logic Elements
Element TypePurposeKey Config
Set ValuesVariable assignment and data transformationelementValueMap (key-value pairs)
ValidationInput validation rules with custom messagesvalidationFormula, errorMessage
FormulaCalculate values using formula expressionsexpression, dataType
Submit ActionFinal submission of collected datapostMessage, preTransformBundle, postTransformBundle

Phase 3: Generation & Validation

# Verify no duplicate Type/SubType/Language exists
sf data query -q "SELECT Id,Name,Type,SubType,Language,IsActive,VersionNumber FROM OmniProcess WHERE Type='<Type>' AND SubType='<SubType>' AND Language='<Language>' AND OmniProcessType='OmniScript'" -o <org>

Build the OmniScript:

  1. Create the OmniProcess record with Type, SubType, Language, and OmniProcessType='OmniScript'
  2. Create OmniProcessElement child records for each Step (Level=0)
  3. Create OmniProcessElement child records for each element within Steps (Level=1+, ordered by Order field)
  4. Configure PropertySetConfig JSON for each element
  5. Wire action elements to their Integration Procedures / DataRaptors

Validation (STRICT MODE):

  • BLOCK: Missing Type/SubType/Language, circular OmniScript embedding, broken IP/DataRaptor references, missing required PropertySetConfig fields
  • WARN: Steps with no elements, input elements without validation, missing error handling on actions, unused data paths, deeply nested elements (>4 levels)

Validation Report Format (6-Category Scoring 0-120):

Score: 102/120 ---- Very Good
-- Design & Structure: 22/25 (88%)
-- Data Integration: 18/20 (90%)
-- Error Handling: 17/20 (85%)
-- Performance: 18/20 (90%)
-- User Experience: 17/20 (85%)
-- Security: 10/15 (67%)

Phase 4: Deployment

  1. Deploy all dependencies first: DataRaptors, Integration Procedures, referenced OmniScripts
  2. Retrieve or deploy OmniScript metadata: sf project retrieve start -m OmniScript:<Name> -o <org> sf project deploy start -m OmniScript:<Name> -o <org>
  3. Activate the OmniScript version after successful deployment
  4. Verify activation: sf data query -q "SELECT Id,Name,Type,SubType,Language,IsActive,VersionNumber FROM OmniProcess WHERE Type='<Type>' AND SubType='<SubType>' AND OmniProcessType='OmniScript' AND IsActive=true" -o <org>

Phase 5: Testing

Walk through all paths with various data scenarios:

  • Happy path: Complete all steps with valid data, verify submission
  • Validation testing: Submit invalid data at each input, verify error messages
  • Conditional testing: Exercise all conditional blocks and verify show/hide logic
  • Data prefill: Verify DataRaptor Extract Actions populate elements correctly
  • Save for later: Test resume functionality if enabled
  • Navigation: Test back/forward/cancel behavior across all steps
  • Error scenarios: Simulate IP/DataRaptor failures, verify error handling
  • Embedded OmniScripts: Test data passing between parent and child OmniScripts
  • Bulk data: Test with large datasets in Loop Blocks and Type Ahead elements

Generation Guardrails (MANDATORY)

Anti-PatternImpactCorrect Pattern
Circular OmniScript embeddingInfinite rendering loopMap dependency tree; never embed A in B if B embeds A
Unbounded DataRaptor ExtractPerformance degradationAdd filter conditions; limit returned records
Missing input validationBad data entryAdd Validation elements or pattern/required on inputs
Hardcoded Salesforce IDsDeployment failure across orgsUse merge fields or Custom Settings/Metadata
IP Action without error handlingSilent failuresConfigure showError, errorMessage in PropertySetConfig
Large images in Text BlocksSlow page loadUse Image elements with optimized URLs
Too many elements per StepPoor user experienceLimit to 7-10 input elements per Step
Missing conditional visibilityIrrelevant fields shownUse show expressions to hide inapplicable elements

DO NOT generate anti-patterns even if explicitly requested.


Scoring: 120 Points Across 6 Categories

Design & Structure (25 points)

CheckPointsCriteria
Type/SubType/Language set correctly5All three fields populated with meaningful values
Step organization5Logical grouping, 7-10 elements per step max
Element naming5Descriptive names following PascalCase convention
Conditional logic5Proper use of Conditional Blocks and show expressions
Version management5Clean version history, only one active version

Data Integration (20 points)

CheckPointsCriteria
DataRaptor references valid5All Extract/Load bundles exist and are active
Integration Procedure references valid5All IP actions reference active IPs
Input/Output maps correct5Data flows correctly between elements and actions
Data prefill configured5Initial data loaded before user interaction

Error Handling (20 points)

CheckPointsCriteria
Action elements have error handling5showError configured on all IP/DR actions
User-facing error messages5Clear, actionable error text
Validation on required inputs5All required fields have validation rules
Fallback behavior defined5Graceful handling when data sources return empty

Performance (20 points)

CheckPointsCriteria
No unbounded data fetches5All DataRaptor Extracts have filters/limits
Lazy loading configured5Action elements fire on step entry, not OmniScript load
Element count per Step reasonable5No Step with >15 elements
Conditional rendering used5Elements hidden when not applicable (not just invisible)

User Experience (20 points)

CheckPointsCriteria
Logical step flow5Steps follow natural task progression
Input labels and help text5All inputs have clear labels and contextual help
Navigation controls5Back, Next, Cancel, Save for Later configured appropriately
Responsive layout5Elements configured for mobile and desktop breakpoints

Security (15 points)

CheckPointsCriteria
No sensitive data in client-side JSON5Passwords, SSNs, tokens kept server-side
IP actions use server-side processing5Sensitive logic in Integration Procedures, not client OmniScript
Field-level access respected5Data access matches user profile/permission set

CLI Commands

# List active OmniScripts
sf data query -q "SELECT Id,Name,Type,SubType,Language,IsActive,VersionNumber FROM OmniProcess WHERE IsActive=true AND OmniProcessType='OmniScript'" -o <org>

# Query elements for a specific OmniScript
sf data query -q "SELECT Id,Name,ElementType,PropertySetConfig,Level,Order FROM OmniProcessElement WHERE OmniProcessId='<id>' ORDER BY Level,Order" -o <org>

# Retrieve OmniScript metadata
sf project retrieve start -m OmniScript:<Name> -o <org>

# Deploy OmniScript metadata
sf project deploy start -m OmniScript:<Name> -o <org>

# Check OmniScript versions
sf data query -q "SELECT Id,VersionNumber,IsActive,LastModifiedDate FROM OmniProcess WHERE Type='<Type>' AND SubType='<SubType>' AND OmniProcessType='OmniScript' ORDER BY VersionNumber DESC" -o <org>

Cross-Skill Integration

From SkillTo sf-omniscriptWhen
sf-omnistudio-analyze-> sf-omniscript"Analyze dependencies before building OmniScript"
sf-datamapper-> sf-omniscript"DataRaptor ready, build the OmniScript that uses it"
sf-integration-procedure-> sf-omniscript"IP ready, wire it into the OmniScript action"
From sf-omniscriptTo SkillWhen
sf-omniscript-> sf-flexcard"Build FlexCard that launches this OmniScript"
sf-omniscript-> sf-deploy"Deploy OmniScript to target org"
sf-omniscript-> sf-omnistudio-analyze"Map full dependency tree before deployment"
sf-omniscript-> sf-integration-procedure"Need a new IP for this OmniScript action"
sf-omniscript-> sf-datamapper"Need a DataRaptor for data prefill"

Edge Cases

ScenarioSolution
Multi-language OmniScriptCreate separate versions per Language with shared Type/SubType. Use translation workbench for labels
Embedded OmniScript data passingMap parent data JSON keys to child OmniScript input via prefillJSON. Test data round-trip
Large Loop Block datasetsPaginate or limit DataRaptor results. Consider server-side filtering in IP
OmniScript in FlexCard flyoutEnsure FlexCard passes required context data. Test flyout sizing
Community/Experience Cloud deploymentVerify OmniScript component is available in Experience Builder. Check guest user permissions
Save & Resume (Save for Later)Configure saveNameTemplate, saveExpireInDays. Test resume with partial data
Versioning conflictsDeactivate old version before activating new. Never have two active versions for same triplet
Custom Lightning Web Components in OmniScriptRegister LWC as OmniScript-compatible. Follow omniscript-lwc namespace conventions

Debug: OmniScript not rendering -> check activation status + element hierarchy | Data not prefilling -> verify DataRaptor Extract output mapping + JSON path | IP action failing -> check IP independently first + verify input map | Steps not showing -> review conditional visibility expressions


Notes

Dependencies (required): sf-datamapper, sf-integration-procedure | Dependencies (optional): sf-deploy, sf-flexcard, sf-omnistudio-analyze | API: 66.0 | Mode: Strict (warnings block) | Scoring: Block deployment if score < 67 | Reference docs: See references/ for element types and best practices

Creating OmniScripts programmatically: Use REST API (sf api request rest --method POST --body @file.json). Required fields: Name, Type, SubType, Language, VersionNumber. OmniScripts default to IsIntegrationProcedure=false (do NOT set OmniProcessType — it is computed). The sf data create record --values flag cannot handle JSON textarea fields like PropertySetConfig. Create child OmniProcessElement records via REST API for each Step and element.

平台分布

Codex

37.55%
按下载量换算42

Claude

29.74%
按下载量换算33

Cursor

19.62%
按下载量换算22

Gemini CLI

9.06%
按下载量换算10

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills