FORMWEFT01 Choose a real problem
For example, organizing orders, analyzing reviews, creating voiceover content, or turning articles into searchable material. Each resource indicates who it's for, what input it requires, and what it delivers.
FORMWEFT RESOURCE LAB
Give your team an AI methodology they can follow.
From order analysis, content creation, knowledge bases, to process handoffs: start with a real problem, learn a method, then use templates to produce your own results.
Don't try to learn all the tools first; complete one thing first.
Download the free learning pack, work through a fictional sample, then decide which capabilities are worth integrating into your business.
02 Ten-minute startup
After extracting, open SKILL.md and provide it along with support files like references/ and templates/ to your AI tool. First, ask it to run only the teaching sample within the pack, then verify against the expected output.
03 Evaluate value based on results
Keep sources, definitions, and review records. Producing a table or draft doesn't mean the system is deployed; connecting real accounts and publishing externally are separate steps.
After learning, start working directly.
Beyond the method packs, there are also original mini-tools you can use in your browser.
Voiceover Motion Workbench
Import local videos and subtitles, drag cards, customize backgrounds, and freely disable watermarks. Export the final video with original audio directly; use MP4 when the browser supports it, otherwise clearly output WebM. Assets stay on your device.
See tools and examplesMake a voiceover first
No account needed, load the sample to practice. Adjust subtitle position, color, and pace first, then export the short video to review; projects can be saved and continued later.
Open the workbench directlyAUDITED RESOURCES
Look at the problem first, then choose the Skill.
23 standalone method packs covering content, business data, development, and enterprise processes. Look at the input and output first, then download and practice; for continuous use, connect the method into a workbench.
Have your own proven method?Share a reproducible Skill to help other teams avoid pitfalls.
Submit a Skill27 public resources in total
No matching resources. Try fewer keywords or select “All resources.”
01Workflow
Content growth and video productionThis version has been reviewed.
Talking-head motion graphics workbench: turn subtitles into visuals
Free voiceover dynamic effects workbench: import videos and SRT locally, drag captions, customize card backgrounds, watermark can be turned off, and export MP4 or WebM output with original audio. Offline source code and Skill included, no asset upload required.
Suitability & requirements
- Best for
- Voiceover creators, enterprise content teams, video editors, and FDE learners
- Reading & practice reference
- 20 minutes
- Version / License
- 1.5.0 · MIT
- Online
- Offline by default
Talking-head motion graphics workbench: turn subtitles into visuals
Free voiceover dynamic effects workbench: import videos and SRT locally, drag captions, customize card backgrounds, watermark can be turned off, and export MP4 or WebM output with original audio. Offline source code and Skill included, no asset upload required.
- Who it's for
- Voiceover creators, enterprise content teams, video editors, and FDE learners
- Reading & practice reference
- 20 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- In-browser local compositing; no dependency installation; supports final videos with original audio and transparent layer export
- Version & license
- 1.5.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/05 · Original browser tool with accompanying documentation. Verified input limits, local asset processing, no external dependencies, and export boundaries; no third-party project code or assets included. Before using your own materials, still confirm authorization and sensitivity levels.
02Workflow
Content growth and video productionThis version has been reviewed.
MiniMax H3 local video workflow: generate raw footage, edit subtitles, export final video
Learn to use local ComfyUI and MiniMax H3 to create voiceover raw footage, then drag captions, customize card backgrounds, and export the final video in the FORMWEFT workbench. Provides original Skill, de-identified workflow, prompts, and caption samples, no model weights included.
Suitability & requirements
- Best for
- Voiceover creators, enterprise content teams, and local AI video learners
- Reading & practice reference
- 35 minutes
- Version / License
- 1.0.0 · MIT (original tutorials and tools); models separately licensed
- Online
- On-demand online
MiniMax H3 local video workflow: generate raw footage, edit subtitles, export final video
Learn to use local ComfyUI and MiniMax H3 to create voiceover raw footage, then drag captions, customize card backgrounds, and export the final video in the FORMWEFT workbench. Provides original Skill, de-identified workflow, prompts, and caption samples, no model weights included.
- Who it's for
- Voiceover creators, enterprise content teams, and local AI video learners
- Reading & practice reference
- 35 minutes; excludes installation, model generation, and business validation
- Internet access
- On-demand online
- Local software / dependencies
- Offline verification script in the package; generation requires separate ComfyUI, Python, GPU, and licensed models
- Version & license
- 1.0.0 · MIT (original tutorials and tools); models are separately licensed
- Public review record
- This version has been reviewed. · 2026/09/05 · Based on one real local H3 generation and voiceover workbench export, this is an original teaching pack. Checked fixed file list, sanitized workflow, offline verification, and subtitle handover; contains no model weights, generated videos, or auto-installers. The sample SRT is an editing script pending audio calibration, not accurate transcription.
03Agent Skill
AI video and content productionThis version has been reviewed.
AI Short-Drama Production Design and Acceptance Skill
Original AI short-drama production design and acceptance Skill: outputs planning, characters, storyboards, task lists, and acceptance sheets by default; only after tools, data destinations, and authorizations are approved does actual generation begin.
Suitability & requirements
- Best for
- Content teams, brand teams, AI video producers, and FDE engineers
- Reading & practice reference
- 45 minutes
- Version / License
- 1.0.0 · MIT
- Online
- Offline by default
AI Short-Drama Production Design and Acceptance Skill
Original AI short-drama production design and acceptance Skill: outputs planning, characters, storyboards, task lists, and acceptance sheets by default; only after tools, data destinations, and authorizations are approved does actual generation begin.
- Who it's for
- Content teams, brand teams, AI video producers, and FDE engineers
- Reading & practice reference
- 45 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- None; approved local or cloud Providers can be connected per project
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/05 · FORMWEFT original documentation-type Skill, intended by default for production planning and acceptance design. Structural checks, offline static scans, and manual file-by-file reviews are complete. The download package contains no scripts, dependencies, model weights, keys, auto-uploads, or auto-publishing. MIT covers only original documentation and templates, not third-party tools, models, assets, or output.
04Agent Skill
AI video and content productionThis version has been reviewed.
AI talking-head and digital human production design Skill
Original AI voiceover and digital human production design Skill: covers licensing, spoken scripts, performance, lip sync, captions, and acceptance; by default does not call models, upload assets, or auto-publish.
Suitability & requirements
- Best for
- Marketing, training, product education, digital IP, and content operations teams
- Reading & practice reference
- 35 minutes
- Version / License
- 1.0.0 · MIT
- Online
- Offline by default
AI talking-head and digital human production design Skill
Original AI voiceover and digital human production design Skill: covers licensing, spoken scripts, performance, lip sync, captions, and acceptance; by default does not call models, upload assets, or auto-publish.
- Who it's for
- Marketing, training, product education, digital IP, and content operations teams
- Reading & practice reference
- 35 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- None; approved TTS, digital human, or editing Providers may be connected per project
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/05 · FORMWEFT original document-type Skill; by default intended only for production planning and acceptance design. Structure validation, offline static scan, and per-file manual review completed; the download package contains no scripts, dependencies, account keys, voice cloning, or auto-publishing. The MIT license covers only original documents and templates, not third-party tools, models, assets, or outputs.
05Checklist
Security, governance, and trusted deploymentThis version has been reviewed.
GitHub repo quick diligence Skill
Without cloning, installing, or running repository code, verify license, maintenance, supply chain, release integrity, and adoption risks to form a traceable GitHub repository due diligence evidence package.
Suitability & requirements
- Best for
- FDE, platform engineering, open-source governance, security, and technology procurement teams
- Reading & practice reference
- 30 minutes
- Version / License
- 1.0.0 · MIT
- Online
- On-demand online
GitHub repo quick diligence Skill
Without cloning, installing, or running repository code, verify license, maintenance, supply chain, release integrity, and adoption risks to form a traceable GitHub repository due diligence evidence package.
- Who it's for
- FDE, platform engineering, open-source governance, security, and technology procurement teams
- Reading & practice reference
- 30 minutes; excludes installation, model generation, and business validation
- Internet access
- On-demand online
- Local software / dependencies
- None; if online verification is needed, the user must approve sites and scope
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/05 · Written independently clean-room based on directory topics; no licensed Feishu original package copied or redistributed. The pinned ZIP has passed structural validation, offline static scanning, and per-file manual review; no scripts, dependencies, install hooks, mandatory network access, or automatic remote operations.
06Agent Skill
Enterprise AI and FDEThis version has been reviewed.
GitHub high-potential project radar Skill
Use a repeatable evidence framework to discover and compare high-potential GitHub projects, distinguishing popularity, maintenance quality, adoption fit, and supply chain risks without treating star counts as conclusions.
Suitability & requirements
- Best for
- FDE, R&D efficiency, technology strategy, innovation, and product research teams
- Reading & practice reference
- 30 minutes
- Version / License
- 1.0.0 · MIT
- Online
- On-demand online
GitHub high-potential project radar Skill
Use a repeatable evidence framework to discover and compare high-potential GitHub projects, distinguishing popularity, maintenance quality, adoption fit, and supply chain risks without treating star counts as conclusions.
- Who it's for
- FDE, R&D efficiency, technology strategy, innovation, and product research teams
- Reading & practice reference
- 30 minutes; excludes installation, model generation, and business validation
- Internet access
- On-demand online
- Local software / dependencies
- None; if public web retrieval is needed, user must approve sites and scope
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/05 · Written independently clean-room based on catalog themes; did not copy or redistribute unlicensed Feishu original package. Fixed ZIP has passed structural validation, offline static scanning, and file-by-file manual review; default read-only, no cloning, running, modifying remote, or auto-contacting authors.
07Agent Skill
E-commerce visuals and content productionThis version has been reviewed.
E-commerce image prompt reconstruction Skill
Break a reference image into composition, subject, lighting, materials, text safe area, and constraints to form reviewable, iterable e-commerce image prompts without directly copying brand assets.
Suitability & requirements
- Best for
- Business operations, marketing, e-commerce, content, data, and FDE teams
- Reading & practice reference
- 25 minutes
- Version / License
- 1.0.0 · MIT
- Online
- Offline by default
E-commerce image prompt reconstruction Skill
Break a reference image into composition, subject, lighting, materials, text safe area, and constraints to form reviewable, iterable e-commerce image prompts without directly copying brand assets.
- Who it's for
- Business operations, marketing, e-commerce, content, data, and FDE teams
- Reading & practice reference
- 25 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- None; only processes reference materials explicitly provided and authorized by the user
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/04 · Rewritten independently based on the Feishu directory topic clean-room, without copying or redistributing the original package text. Conclusions apply only to the current fixed hash: the archive structure has been parsed, static rule result is eligible, with 6 files and no scripts, dependencies, or install hooks; online boundaries follow the statements in this resource, and files are output only to user-approved locations. Static review cannot replace runtime environment, input data, or later-version rechecks.
08Agent Skill
E-commerce visuals and content productionThis version has been reviewed.
E-commerce main image evidence-based diagnosis Skill
Diagnose e-commerce product images using five dimensions: information hierarchy, product recognition, selling points, platform constraints, and experimental hypotheses, and output actionable change orders rather than just aesthetic opinions.
Suitability & requirements
- Best for
- Business operations, marketing, e-commerce, content, data, and FDE teams
- Reading & practice reference
- 30 minutes
- Version / License
- 1.0.0 · MIT
- Online
- Offline by default
E-commerce main image evidence-based diagnosis Skill
Diagnose e-commerce product images using five dimensions: information hierarchy, product recognition, selling points, platform constraints, and experimental hypotheses, and output actionable change orders rather than just aesthetic opinions.
- Who it's for
- Business operations, marketing, e-commerce, content, data, and FDE teams
- Reading & practice reference
- 30 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- None; use a browser or image viewer to open user-approved materials as needed.
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/04 · Rewritten independently based on the Feishu directory topic clean-room, without copying or redistributing the original package text. Conclusions apply only to the current fixed hash: the archive structure has been parsed, static rule result is eligible, with 6 files and no scripts, dependencies, or install hooks; online boundaries follow the statements in this resource, and files are output only to user-approved locations. Static review cannot replace runtime environment, input data, or later-version rechecks.
09Agent Skill
E-commerce operations and data analysisThis version has been reviewed.
E-commerce operations data analysis Skill
Establish a unified standard from orders, traffic, ads, products, and fulfillment data to identify profit and growth issues, forming a business analysis with evidence, responsible parties, and next steps.
Suitability & requirements
- Best for
- Business operations, marketing, e-commerce, content, data, and FDE teams
- Reading & practice reference
- 40 minutes
- Version / License
- 1.0.0 · MIT
- Online
- Offline by default
E-commerce operations data analysis Skill
Establish a unified standard from orders, traffic, ads, products, and fulfillment data to identify profit and growth issues, forming a business analysis with evidence, responsible parties, and next steps.
- Who it's for
- Business operations, marketing, e-commerce, content, data, and FDE teams
- Reading & practice reference
- 40 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- None; you may use company-approved spreadsheet tools to process local copies
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/04 · Rewritten independently based on the Feishu directory topic clean-room, without copying or redistributing the original package text. Conclusions apply only to the current fixed hash: the archive structure has been parsed, static rule result is eligible, with 6 files and no scripts, dependencies, or install hooks; online boundaries follow the statements in this resource, and files are output only to user-approved locations. Static review cannot replace runtime environment, input data, or later-version rechecks.
10Agent Skill
Automation and workflowsThis version has been reviewed.
n8n content pipeline blueprint Skill
First clarify the boundaries of input, approval, generation, quality inspection, publishing, and rollback, then configure the content production workflow into an auditable n8n blueprint, defaulting to not connecting any external accounts.
Suitability & requirements
- Best for
- Business operations, marketing, e-commerce, content, data, and FDE teams
- Reading & practice reference
- 35 minutes
- Version / License
- 1.0.0 · MIT
- Online
- On-demand online
n8n content pipeline blueprint Skill
First clarify the boundaries of input, approval, generation, quality inspection, publishing, and rollback, then configure the content production workflow into an auditable n8n blueprint, defaulting to not connecting any external accounts.
- Who it's for
- Business operations, marketing, e-commerce, content, data, and FDE teams
- Reading & practice reference
- 35 minutes; excludes installation, model generation, and business validation
- Internet access
- On-demand online
- Local software / dependencies
- None; optional n8n during implementation, must approve instance, domain, credentials, and permissions first
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/04 · Rewritten independently based on the Feishu directory topic clean-room, without copying or redistributing the original package text. Conclusions apply only to the current fixed hash: the archive structure has been parsed, static rule result is eligible, with 6 files and no scripts, dependencies, or install hooks; online boundaries follow the statements in this resource, and files are output only to user-approved locations. Static review cannot replace runtime environment, input data, or later-version rechecks.
11Agent Skill
E-commerce visuals and content productionThis version has been reviewed.
E-commerce visual production SOP Skill
Connect requirements, assets, layouts, generation, manual review, platform adaptation, and archiving into a transferable e-commerce visual production standard, reducing rework and version chaos.
Suitability & requirements
- Best for
- Business operations, marketing, e-commerce, content, data, and FDE teams
- Reading & practice reference
- 35 minutes
- Version / License
- 1.0.0 · MIT
- Online
- Offline by default
E-commerce visual production SOP Skill
Connect requirements, assets, layouts, generation, manual review, platform adaptation, and archiving into a transferable e-commerce visual production standard, reducing rework and version chaos.
- Who it's for
- Business operations, marketing, e-commerce, content, data, and FDE teams
- Reading & practice reference
- 35 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- None; may use the company's existing design and asset management tools
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/04 · Rewritten independently based on the Feishu directory topic clean-room, without copying or redistributing the original package text. Conclusions apply only to the current fixed hash: the archive structure has been parsed, static rule result is eligible, with 6 files and no scripts, dependencies, or install hooks; online boundaries follow the statements in this resource, and files are output only to user-approved locations. Static review cannot replace runtime environment, input data, or later-version rechecks.
12Agent Skill
E-commerce growth and experimentationThis version has been reviewed.
Viral main image evidence lab Skill
Use samples, variables, evidence, and experiment records to study high-performing images, distinguish correlation from causation, avoid copying competitor assets, and produce verifiable creative hypotheses.
Suitability & requirements
- Best for
- Business operations, marketing, e-commerce, content, data, and FDE teams
- Reading & practice reference
- 45 minutes
- Version / License
- 1.0.0 · MIT
- Online
- On-demand online
Viral main image evidence lab Skill
Use samples, variables, evidence, and experiment records to study high-performing images, distinguish correlation from causation, avoid copying competitor assets, and produce verifiable creative hypotheses.
- Who it's for
- Business operations, marketing, e-commerce, content, data, and FDE teams
- Reading & practice reference
- 45 minutes; excludes installation, model generation, and business validation
- Internet access
- On-demand online
- Local software / dependencies
- None; before studying external samples, must approve specific public domains and recording scope.
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/04 · Rewritten independently based on the Feishu directory topic clean-room, without copying or redistributing the original package text. Conclusions apply only to the current fixed hash: the archive structure has been parsed, static rule result is eligible, with 6 files and no scripts, dependencies, or install hooks; online boundaries follow the statements in this resource, and files are output only to user-approved locations. Static review cannot replace runtime environment, input data, or later-version rechecks.
13Agent Skill
Competitive intelligence and business strategyThis version has been reviewed.
Competitor evidence distillation map Skill
Organize public competitor information into a four-layer map of facts, inferences, unknowns, and actions, retaining sources and timestamps, so competitive intelligence truly serves product, marketing, and sales decisions.
Suitability & requirements
- Best for
- Business operations, marketing, e-commerce, content, data, and FDE teams
- Reading & practice reference
- 50 minutes
- Version / License
- 1.0.0 · MIT
- Online
- On-demand online
Competitor evidence distillation map Skill
Organize public competitor information into a four-layer map of facts, inferences, unknowns, and actions, retaining sources and timestamps, so competitive intelligence truly serves product, marketing, and sales decisions.
- Who it's for
- Business operations, marketing, e-commerce, content, data, and FDE teams
- Reading & practice reference
- 50 minutes; excludes installation, model generation, and business validation
- Internet access
- On-demand online
- Local software / dependencies
- None; public research only after approved domains, terms, and scope.
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/04 · Rewritten independently based on the Feishu directory topic clean-room, without copying or redistributing the original package text. Conclusions apply only to the current fixed hash: the archive structure has been parsed, static rule result is eligible, with 6 files and no scripts, dependencies, or install hooks; online boundaries follow the statements in this resource, and files are output only to user-approved locations. Static review cannot replace runtime environment, input data, or later-version rechecks.
14Agent Skill
Content growth and video productionThis version has been reviewed.
Competitor teardown script toolkit Skill
Turn competitor research into an evidence chain, shot list, and review checklist for short videos or talking-head scripts, avoiding unverified comparisons, disparagement, and copyright material reuse.
Suitability & requirements
- Best for
- Business operations, marketing, e-commerce, content, data, and FDE teams
- Reading & practice reference
- 45 minutes
- Version / License
- 1.0.0 · MIT
- Online
- On-demand online
Competitor teardown script toolkit Skill
Turn competitor research into an evidence chain, shot list, and review checklist for short videos or talking-head scripts, avoiding unverified comparisons, disparagement, and copyright material reuse.
- Who it's for
- Business operations, marketing, e-commerce, content, data, and FDE teams
- Reading & practice reference
- 45 minutes; excludes installation, model generation, and business validation
- Internet access
- On-demand online
- Local software / dependencies
- None; specific domains must be approved before verifying public sources, and production tools are approved separately by the business.
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/04 · Rewritten independently based on the Feishu directory topic clean-room, without copying or redistributing the original package text. Conclusions apply only to the current fixed hash: the archive structure has been parsed, static rule result is eligible, with 6 files and no scripts, dependencies, or install hooks; online boundaries follow the statements in this resource, and files are output only to user-approved locations. Static review cannot replace runtime environment, input data, or later-version rechecks.
15Agent Skill
Automation and workflowsThis version has been reviewed.
Feishu data hub: reconcile first, then import
Organize order, product, and refund CSV files into clear field definitions, primary keys, currencies, and import previews, avoiding writing dirty data straight into online spreadsheets.
Suitability & requirements
- Best for
- E-commerce operations, finance, data leads, and project teams preparing to build Feishu multidimensional tables
- Reading & practice reference
- 40 minutes
- Version / License
- 1.0.0 · MIT
- Online
- Offline by default
Feishu data hub: reconcile first, then import
Organize order, product, and refund CSV files into clear field definitions, primary keys, currencies, and import previews, avoiding writing dirty data straight into online spreadsheets.
- Who it's for
- E-commerce operations, finance, data leads, and project teams preparing to build Feishu multidimensional tables
- Reading & practice reference
- 40 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- None; documents, templates, and samples
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/05 · FORMWEFT independently authored Chinese methods, templates, and fictional samples; Feishu original package not redistributed. Current fixed version has completed file structure, static rules, and content review, no high-risk behavior found. Real data permissions and generated results still need user review.
16Agent Skill
Automation and workflowsThis version has been reviewed.
Article knowledge base: every answer has a source
Turn authorized articles and internal documents into a source ledger, knowledge chunks, retrieval questions, and answer evidence, handling outdated versions, conflicts, and insufficient materials.
Suitability & requirements
- Best for
- Knowledge operations, customer service, training, policy management, and enterprise AI project leads
- Reading & practice reference
- 35 minutes
- Version / License
- 1.0.0 · MIT
- Online
- Offline by default
Article knowledge base: every answer has a source
Turn authorized articles and internal documents into a source ledger, knowledge chunks, retrieval questions, and answer evidence, handling outdated versions, conflicts, and insufficient materials.
- Who it's for
- Knowledge operations, customer service, training, policy management, and enterprise AI project leads
- Reading & practice reference
- 35 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- None; documents, templates, and samples
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/05 · FORMWEFT independently authored Chinese methods, templates, and fictional samples; Feishu original package not redistributed. Current fixed version has completed file structure, static rules, and content review, no high-risk behavior found. Real data permissions and generated results still need user review.
17Agent Skill
Content growth and video productionThis version has been reviewed.
Tool hands-on reviews: no steps left to imagination
Write reproducible tutorials using real versions, operation logs, screenshots, and read-back evidence; parts not actually verified remain as pending verification rather than being fabricated as definite steps.
Suitability & requirements
- Best for
- Corporate training, customer success, product operations, technical writing, and internal tool leads
- Reading & practice reference
- 30 minutes
- Version / License
- 1.0.0 · MIT
- Online
- Offline by default
Tool hands-on reviews: no steps left to imagination
Write reproducible tutorials using real versions, operation logs, screenshots, and read-back evidence; parts not actually verified remain as pending verification rather than being fabricated as definite steps.
- Who it's for
- Corporate training, customer success, product operations, technical writing, and internal tool leads
- Reading & practice reference
- 30 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- None; documents, templates, and samples
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/05 · FORMWEFT independently authored Chinese methods, templates, and fictional samples; Feishu original package not redistributed. Current fixed version has completed file structure, static rules, and content review, no high-risk behavior found. Real data permissions and generated results still need user review.
18Agent Skill
Content growth and video productionThis version has been reviewed.
Xiaohongshu content planning: from evidence to review
Score topics, create image-text drafts, hand off for manual publishing, and review performance based on real audience questions and brand evidence, with account actions always individually authorized.
Suitability & requirements
- Best for
- Brand marketing, content operations, founder IP teams, and enterprise content leads
- Reading & practice reference
- 35 minutes
- Version / License
- 1.0.0 · MIT
- Online
- Offline by default
Xiaohongshu content planning: from evidence to review
Score topics, create image-text drafts, hand off for manual publishing, and review performance based on real audience questions and brand evidence, with account actions always individually authorized.
- Who it's for
- Brand marketing, content operations, founder IP teams, and enterprise content leads
- Reading & practice reference
- 35 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- None; documents, templates, and samples
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/05 · FORMWEFT independently authored Chinese methods, templates, and fictional samples; Feishu original package not redistributed. Current fixed version has completed file structure, static rules, and content review, no high-risk behavior found. Real data permissions and generated results still need user review.
19Agent Skill
Enterprise AI and FDEThis version has been reviewed.
UI design review: turn 'premium feel' into testable questions
Use page goals, information hierarchy, font size and line height, spacing and alignment, and responsive testing to turn vague aesthetic opinions into a corrective checklist with evidence, priorities, and re-testing conditions.
Suitability & requirements
- Best for
- Product leads, UI/UX designers, frontend engineers, and FDE teams involved in delivery
- Reading & practice reference
- 30 minutes
- Version / License
- 1.0.0 · MIT
- Online
- Offline by default
UI design review: turn 'premium feel' into testable questions
Use page goals, information hierarchy, font size and line height, spacing and alignment, and responsive testing to turn vague aesthetic opinions into a corrective checklist with evidence, priorities, and re-testing conditions.
- Who it's for
- Product leads, UI/UX designers, frontend engineers, and FDE teams involved in delivery
- Reading & practice reference
- 30 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- None; documents, templates, and samples
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/05 · FORMWEFT independently authored Chinese methods, templates, and fictional samples; Feishu original package not redistributed. Current fixed version has completed file structure, static rules, and content review, no high-risk behavior found. Real data permissions and generated results still need user review.
20Agent Skill
E-commerce operations and data analysisThis version has been reviewed.
Customer review analysis: dedupe first, then discuss trends
Define deduplication, tag definitions, unique counts, denominators, and priorities clearly to avoid duplicate reviews, word frequency, and a few quotes being mistaken for all customer conclusions.
Suitability & requirements
- Best for
- Product, customer service, e-commerce operations, customer success, quality, and enterprise data teams
- Reading & practice reference
- 35 minutes
- Version / License
- 1.0.0 · MIT
- Online
- Offline by default
Customer review analysis: dedupe first, then discuss trends
Define deduplication, tag definitions, unique counts, denominators, and priorities clearly to avoid duplicate reviews, word frequency, and a few quotes being mistaken for all customer conclusions.
- Who it's for
- Product, customer service, e-commerce operations, customer success, quality, and enterprise data teams
- Reading & practice reference
- 35 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- None; documents, templates, and samples
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/05 · FORMWEFT independently authored Chinese methods, templates, and fictional samples; Feishu original package not redistributed. Current fixed version has completed file structure, static rules, and content review, no high-risk behavior found. Real data permissions and generated results still need user review.
21Agent Skill
Content growth and video productionThis version has been reviewed.
TikTok content operations: plan that executes, account without overreach
Use brand evidence to select original topics, organize weekly content, and vet creators; reserve login, contact, publishing, ad placement, and budget strictly for clearly authorized personnel.
Suitability & requirements
- Best for
- Overseas brands, content operations, short-video production, creator partnerships, and growth teams
- Reading & practice reference
- 35 minutes
- Version / License
- 1.0.0 · MIT
- Online
- Offline by default
TikTok content operations: plan that executes, account without overreach
Use brand evidence to select original topics, organize weekly content, and vet creators; reserve login, contact, publishing, ad placement, and budget strictly for clearly authorized personnel.
- Who it's for
- Overseas brands, content operations, short-video production, creator partnerships, and growth teams
- Reading & practice reference
- 35 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- None; documents, templates, and samples
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/05 · FORMWEFT independently authored Chinese methods, templates, and fictional samples; Feishu original package not redistributed. Current fixed version has completed file structure, static rules, and content review, no high-risk behavior found. Real data permissions and generated results still need user review.
22Agent Skill
E-commerce operations and data analysisThis version has been reviewed.
Amazon operations review: orders, ads, inventory on the same basis
Separate order, refund, ad, and inventory granularity, then use traceable formulas to calculate ACOS, TACOS, refunds, and coverage days, forming pending actions that do not auto-modify accounts.
Suitability & requirements
- Best for
- Amazon operations, advertising, supply chain, finance, brand leads, and data teams
- Reading & practice reference
- 40 minutes
- Version / License
- 1.0.0 · MIT
- Online
- Offline by default
Amazon operations review: orders, ads, inventory on the same basis
Separate order, refund, ad, and inventory granularity, then use traceable formulas to calculate ACOS, TACOS, refunds, and coverage days, forming pending actions that do not auto-modify accounts.
- Who it's for
- Amazon operations, advertising, supply chain, finance, brand leads, and data teams
- Reading & practice reference
- 40 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- None; documents, templates, and samples
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/05 · FORMWEFT independently authored Chinese methods, templates, and fictional samples; Feishu original package not redistributed. Current fixed version has completed file structure, static rules, and content review, no high-risk behavior found. Real data permissions and generated results still need user review.
23Agent Skill
Data analysis and dashboardsThis version has been reviewed.
Group chat insights: see problems, don't profile individuals
Turn authorized, exported, minimized group chat spreadsheets into a ledger of group-level issues, decisions, and actions; after deduplication, disclose the denominator, protect privacy, and avoid profiling members by performance or personality.
Suitability & requirements
- Best for
- Business operations, customer success, project leads, knowledge management, and teams needing to review collaboration issues
- Reading & practice reference
- 35 minutes
- Version / License
- 1.0.0 · MIT
- Online
- Offline by default
Group chat insights: see problems, don't profile individuals
Turn authorized, exported, minimized group chat spreadsheets into a ledger of group-level issues, decisions, and actions; after deduplication, disclose the denominator, protect privacy, and avoid profiling members by performance or personality.
- Who it's for
- Business operations, customer success, project leads, knowledge management, and teams needing to review collaboration issues
- Reading & practice reference
- 35 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- None; documents, templates, and samples
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/05 · FORMWEFT independently authored Chinese methods, templates, and fictional samples; Feishu original package not redistributed. Current fixed version has completed file structure, static rules, and content review, no high-risk behavior found. Real data permissions and generated results still need user review.
24Agent Skill
Agents and workflowsThis version has been reviewed.
RPA process blueprint: design for failure first, then discuss automation
Break processes into deterministic, assisted, and fully manual steps; define states, idempotency, authoritative readback, failure paths, and takeover materials; stop immediately at CAPTCHAs - do not auto-recognize, bypass, or retry.
Suitability & requirements
- Best for
- Business operations, process leads, automation engineers, risk control, operations, and teams preparing to adopt RPA
- Reading & practice reference
- 40 minutes
- Version / License
- 1.0.0 · MIT
- Online
- Offline by default
RPA process blueprint: design for failure first, then discuss automation
Break processes into deterministic, assisted, and fully manual steps; define states, idempotency, authoritative readback, failure paths, and takeover materials; stop immediately at CAPTCHAs - do not auto-recognize, bypass, or retry.
- Who it's for
- Business operations, process leads, automation engineers, risk control, operations, and teams preparing to adopt RPA
- Reading & practice reference
- 40 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- None; documents, templates, and samples
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/05 · FORMWEFT independently authored Chinese methods, templates, and fictional samples; Feishu original package not redistributed. Current fixed version has completed file structure, static rules, and content review, no high-risk behavior found. Real data permissions and generated results still need user review.
25Agent Skill
Enterprise AI and FDEThis version has been reviewed.
Team SOP handover: only counts as received when someone can pass independently
Turn role experience into role-based SOPs, least-privilege checklists, exception paths, and calculable acceptance records; document handoff is not done until independent reproduction and reverse handoff serve as release evidence.
Suitability & requirements
- Best for
- Business managers, FDE engineers, project delivery, customer success, training leads, and teams taking over roles
- Reading & practice reference
- 35 minutes
- Version / License
- 1.0.0 · MIT
- Online
- Offline by default
Team SOP handover: only counts as received when someone can pass independently
Turn role experience into role-based SOPs, least-privilege checklists, exception paths, and calculable acceptance records; document handoff is not done until independent reproduction and reverse handoff serve as release evidence.
- Who it's for
- Business managers, FDE engineers, project delivery, customer success, training leads, and teams taking over roles
- Reading & practice reference
- 35 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- None; documents, templates, and samples
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/05 · FORMWEFT independently authored Chinese methods, templates, and fictional samples; Feishu original package not redistributed. Current fixed version has completed file structure, static rules, and content review, no high-risk behavior found. Real data permissions and generated results still need user review.
26Templates
Enterprise AI and FDEThis version has been reviewed.
Enterprise AI practice collection: 23 methods in one take
Combine R&D, short-drama voiceover, e-commerce, data, Feishu, knowledge base, and process handoff methods into one learning collection. Choose a set based on business problems, do a sample first, then apply to your own materials.
Suitability & requirements
- Best for
- Business teams and FDE learners who want to practice AI with real business
- Reading & practice reference
- 15 minutes
- Version / License
- 1.0.0 · MIT
- Online
- Offline by default
Enterprise AI practice collection: 23 methods in one take
Combine R&D, short-drama voiceover, e-commerce, data, Feishu, knowledge base, and process handoff methods into one learning collection. Choose a set based on business problems, do a sample first, then apply to your own materials.
- Who it's for
- Business teams and FDE learners who want to practice AI with real business
- Reading & practice reference
- 15 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- None; documents, templates, and samples
- Version & license
- 1.0.0 · MIT
- Public review record
- This version has been reviewed. · 2026/09/05 · FORMWEFT independently authored Chinese methods, templates, and fictional samples; Feishu original package not redistributed. Current fixed version has completed file structure, static rules, and content review, no high-risk behavior found. Real data permissions and generated results still need user review.
27Checklist
Security, governance, and trusted deploymentConstrained learning
Agent Skill security audit and clean release guide
Learn how to isolate, statically check, verify licenses, declare capabilities, manually review, and cleanly rebuild third-party Agent Skills to avoid installing unknown code directly into production environments.
Suitability & requirements
- Best for
- Enterprise AI platforms, security, IT, open-source governance, and resource maintenance personnel
- Reading & practice reference
- 30 minutes
- Version / License
- 0.9.0 · MIT
- Online
- Offline by default
View full methodNo direct download provided
Agent Skill security audit and clean release guide
Learn how to isolate, statically check, verify licenses, declare capabilities, manually review, and cleanly rebuild third-party Agent Skills to avoid installing unknown code directly into production environments.
- Who it's for
- Enterprise AI platforms, security, IT, open-source governance, and resource maintenance personnel
- Reading & practice reference
- 30 minutes; excludes installation, model generation, and business validation
- Internet access
- Offline by default
- Local software / dependencies
- Python 3 standard library scanner (download not yet open)
- Version & license
- 0.9.0 · MIT
- Public review record
- Constrained learning · 2026/09/05 · Methods and rules are public; scanner source code still requires independent code review and malicious archive fixture validation; currently for learning only, no executable package download.
Share your validated methods with the next practitioner.
Welcome submissions of AI Skills that solve real problems. Provide reproducible steps, examples, and usage boundaries. Quality works will be credited and included after review.
24 original packages reviewed; only independently rebuilt versions are released publicly.
Original attachments fully obtained and isolated. Risks in execution, networking, CAPTCHA automation, privacy, and authorization were found, so original attachments were not released directly. The study packages below are rewritten from a business-topic perspective.
24 / 24 obtained
Recorded per-package source, attachment structure, and fixed hash.
24 / 24 static review completed
Including nested documents and script recheck; original packages were not installed or executed. Static review cannot prove absence of backdoors.
23 independent study packages
Covering R&D, content, e-commerce, data, Feishu, knowledge base, and process handover; single and bundle downloads provided.
Publicly learnable and adaptable
MIT covers only documents, templates, and examples written by FORMWEFT; it does not authorize use of third-party brands, materials, models, or customer data.
FROM RESOURCE TO CAPABILITY
Download is not the goal; team capability is.
Verify resources in a real task, with FDE engineers completing configuration, evidence recording, controlled rollout, and internal handover.
Understand the applicable problem first
Use resource detail pages to judge inputs, outputs, boundaries, and non-applicable cases; do not create needs for the sake of tools.
- Stay
- Problem definition and applicability assessment
Practice in a non-sensitive data environment
Run the full process with sample data, recording dependencies, permissions, failure states, and manual confirmation points.
- Stay
- Practice records and risk checklist
Integrate one real business process
Limit user, data, and action scope; validate quality, latency, cost, and rollback with representative samples.
- Stay
- Pilot version, evaluation, and stop conditions
Enable the business team to take over
Complete the runbook, knowledge updates, version tracking, and review so internal staff can continue maintenance.
- Stay
- Operational evidence and capability archive
Review is about this version, not just a “safe” label.
Source, license, capability claims, static rules, human review, version, and SHA-256 must be visible. For third-party packages without redistribution license, only original source and research conclusions are kept.
Isolate
Do not install, execute, or read user credentials.
Inventory
List files, dependencies, network, processes, and write capabilities.
Review
After rule checks, humans still read the code and licenses.
Clean release
Repackage with fixed version, hash, and known limitations.
Incubate your workflow into a Skill your team can use.
Describe the current task, existing process, data and tools, and the deliverables you want to keep. We can organize it into a teachable, testable, maintainable enterprise capability package.