qfeius/make-platform-skills

make-integration

Use when the user asks to recognize invoices, bills, receipts, tickets, or other voucher files with Make Integration services.

Ver código fuente

Cómo usar

  1. Copia el comando y revisa el paquete antes de instalarlo en tu entorno de agente.
  2. Consulta las herramientas, credenciales y entornos necesarios en el documento original.
  3. Describe las entradas y el resultado deseado; contrasta la salida con las fuentes.
npx skills add https://github.com/qfeius/make-platform-skills --skill make-integration

Antes de empezar

  • La instalación no incluye suscripción al modelo, cuota API ni permisos sobre tus archivos.
  • Las estrellas e instalaciones reflejan actividad, no una prueba independiente de calidad o seguridad.
Documento original del Skill

Contenido del repositorio de origen con títulos, ejemplos, código, tablas, enlaces e imágenes preservados.

make-integration

When to use

Use this skill when the user needs Make Platform integration capabilities, especially OCR recognition for invoices and bills.

Current capability:

  • Bill and invoice OCR through makecli integration ocr
  • Optional invoice authenticity verification
  • Structured extraction for PDF/OFD/PNG/JPG/JPEG files
  • Returned invoice fields, stamps, QR codes, coordinates, and optional crop images

Dependency

make-integration depends on the makecli skill and the installed makecli binary.

Instructions

Pre-flight check

Before running OCR:

  1. Check makecli is installed.
  • Missing: run brew tap qfeius/makecli && brew install makecli
  1. Check the OCR command exists.
  • Run makecli integration ocr --help
  • If the command is missing, run makecli update or upgrade makecli.
  1. Check the selected profile.
  • Run makecli configure verify --profile <profile> --output=json
  • If not configured, tell the user to run the interactive setup:

! makecli configure token --profile <profile>

  • If the token is invalid or missing, stop and ask the user to configure makecli credentials. Do not recognize, extract, infer, correct, or complete invoice content outside makecli.
  1. Check the request base URL from makecli.
  • Run makecli configure get server-url --profile <profile>
  • Prefer using the profile-configured server URL instead of hard-coding an environment URL.
  • If the profile has no server URL, makecli uses its own default base URL.
  1. Check the input file.
  • Supported extensions: .pdf, .ofd, .png, .jpg, .jpeg
  • Only upload files provided or approved by the user.

Recognition workflow

Use table output when the user only needs a readable answer:

bash
makecli integration ocr -f <file> --profile <profile>

Use JSON output when you need to parse fields or feed the result into another Make operation:

bash
makecli integration ocr -f <file> --output=json --profile <profile>

Common options:

bash
# Recognize selected pages
makecli integration ocr -f invoice.pdf --pages "1,3,2"
makecli integration ocr -f invoice.pdf --pages "2-4"

# Bind a business document id
makecli integration ocr -f invoice.pdf --business-id <business-id>

# Explicitly disable invoice authenticity verification
makecli integration ocr -f invoice.pdf --verify-vat=false

# Optional response controls
makecli integration ocr -f invoice.pdf --coord-restore-original
makecli integration ocr -f invoice.pdf --crop-complete
makecli integration ocr -f invoice.pdf --crop-value
makecli integration ocr -f invoice.pdf --merge-elec
makecli integration ocr -f invoice.pdf --return-ppi

Output handling

For --output=json, summarize from:

text
data.file_name
data.bill_count
data.processing_duration_ms
data.result.pages[].bills[].items[]
data.result.pages[].bills[].stamps[]
data.result.pages[].bills[].qr_codes[]

When reporting results to the user:

  • Include file name, bill count, page count, and processing duration when available.
  • Extract non-empty items[].value fields.
  • Include stamps and QR codes when they are present and useful.
  • Keep base64 crop payloads out of normal summaries.
  • Report recognized OCR values only from successful makecli integration ocr output.
  • Treat only makecli integration ocr output as the recognition result. Do not use visual inspection, model knowledge, file parsing, alternate OCR tools, direct HTTP calls, or user-visible image content to recognize or fill invoice data.
  • Mask or replace special and personal information only in skill documentation examples, committed samples, logs, or debug snippets.
  • If the OCR result will be written to Make records, invoke the makecli skill for record operations after recognition.

Failure handling

  • API 错误 [990300403]: token验证失败
  • The token is invalid for the selected environment. Ask the user to configure a token valid for that environment, then retry OCR through makecli.
  • API 错误 [990300404]: 服务不存在
  • The selected makecli server URL likely has not deployed the OCR route. Check the profile server URL or confirm deployment.
  • unsupported file extension
  • Ask for a supported PDF/OFD/PNG/JPG/JPEG file.
  • 打开文件失败
  • Check the local path and permissions.
  • Any makecli OCR failure
  • Report the makecli failure and the next makecli configuration or environment step. Do not continue by recognizing the invoice content yourself.

Hard rules

  • Do not write user-provided tokens into repository files.
  • Prefer temporary config directories when the user provides a one-off token.
  • Clean up temporary credentials and OCR result files after the run unless the user asks to keep them.
  • Never print, commit, or duplicate tokens.
  • Do not upload local files unless the user provided or approved them.
  • Do not commit real invoice numbers, taxpayer IDs, order numbers, phone numbers, addresses, tokens, QR payloads, or other sensitive values in skill examples.
  • Do not use any non-makecli path as a fallback for this skill. If makecli cannot run or OCR fails, fix makecli configuration/environment first and retry through makecli.

Relacionados por tarea

del mismo repositorio

Más Skills

Todos los Skills
qfeius
Comunidad

make-app-filter

Use when generating, integrating, reviewing, or debugging Make App record-list keyword search through Service filter.expression or advanced/conditional/table/header filtering with @qfei-design/make-app-filter. Triggered by 列表搜索, 关键词搜索, 筛选, 高级筛选, 条件筛选, 表格/表头/列头/按字段筛选, CEL/DNF, field-type operators, candidate values, and filter Preset save/echo. In search-only mode, use the package compiler without a filter panel or Preset lifecycle. In advanced-filter mode, use the package panel, permission-aware

Instalaciones / 1h
1
GitHub Stars
0
Actualizado
10 nov
qfeius
Comunidad

make-app-observability

Use when creating, integrating, reviewing, or debugging a Make App's default Trace ID chain, traceparent and X-Log-Id propagation, safe correlated logs, or @qfei-design/make-app-observability error notices. Every new Make App includes this baseline even when the user does not mention observability. Covers direct-gateway and Service-fronted Apps and AI transports when present. Does not own authentication, Service route design, AI protocol, tracing backend/exporter deployment, runtime packaging, o

Instalaciones / 1h
1
GitHub Stars
0
Actualizado
10 nov
qfeius
Comunidad

make-app-runtime

Use when generating, refactoring, reviewing, or debugging Make App project runtime structure, workspace manifests, Service runtime, local/dev scripts, build outputs, Docker/K8s image entrypoints, publish readiness, or packaging errors such as missing apps/service/dist/server.js. Covers apps/ workspace contracts, apps/ui/dist, apps/service port/build/start contracts, runtime config file location, runtime artifact tests, forwarded host/proto header preservation, and publish gates that include auth

Instalaciones / 1h
1
GitHub Stars
0
Actualizado
10 nov
qfeius
Comunidad

make-app-service

Use when generating, refactoring, reviewing, or debugging Make App apps/service APIs and UI-Service contracts, including server-side aggregation for dashboards, charts and 统计/汇总. Covers route design, apps/docs/api.md, layered structure, Make adapters, direct Make response passthrough, schema normalization including independent fields/createFields collections, record CRUD, record-write-permission and records/bulk, list filter/sort/groupFilter parsing, record groups, Entity Preset, candidate/looku

Instalaciones / 1h
1
GitHub Stars
0
Actualizado
10 nov