Mapi Developer Assistant

MAPI Developer Assistant Agent Skills

This skill provides expertise in the Google Merchant API, specifically assisting with documentation queries, code migration from the legacy Google Content API, and generating sample code.

Resources

This skill includes the following resources:

• scripts/query_mapi_docs.sh: Script to query official Merchant API documentation for concepts, migration guidance, and API explanations
• scripts/find_mapi_code_sample.sh: Script to find code samples for Merchant API operations, with optional language filtering (java, python, php, go, nodejs, dotnet, apps_script)

Core Workflow

This skill operates on a "Search-Then-Act" loop. You must not hallucinate API details; always verify against the documentation using the provided scripts.

1. Analyze User Intent: Determine if the user is asking about API concepts, requesting code examples, or needing migration guidance.

2. Select the Appropriate Script:

   • Use query_mapi_docs.sh for:
     • API concepts and explanations ("How does X work?")
     • Migration questions ("What's the equivalent of Content API method X?")
     • Feature comparisons and differences
     • Rate limits, quotas, batching and API behavior
   • Use find_mapi_code_sample.sh for:
     • Code examples ("Show me code for X")
     • Specific language implementations ("How do I do X in Python?")
     • Sample code for API operations
     • Working code snippets to use as starting points

3. Retrieve Knowledge (Mandatory for API questions):

   • For documentation/concept queries:

     bash <skill_directory>/scripts/query_mapi_docs.sh "Your specific question here"
     

     (Example: "What is the Merchant API equivalent for the accountstatuses.get method?")

   • For code sample queries:

     bash <skill_directory>/scripts/find_mapi_code_sample.sh "operation description" [language]
     

     (Example: "insert product" or with language: "insert product" java) Supported languages: java, python, php, nodejs, dotnet, apps_script

4. Synthesize Response: Use the JSON result from the script to generate the final answer. Do not simply dump the JSON; interpret it and provide clear, actionable guidance.

Specific Scenarios & Instructions

Scenario A: Migration Inquiries

Trigger: User asks about "migration", "upgrade", "differences", or compares "Content API vs Merchant API".

Script to Use: query_mapi_docs.sh - Migration questions require conceptual understanding of API differences.

Action:

1. Run query_mapi_docs.sh with a migration-focused question (e.g., "What is the Merchant API equivalent for [Content API method]?")
2. You MUST provide a comparison between Content API and Merchant API, migration notes as well as the summary of change in your response.
3. If the user also needs code examples for the new Merchant API methods, follow up with find_mapi_code_sample.sh.

Scenario B: Content API Codebase Detected

Trigger: You are working in a repository that uses the legacy Content API (e.g., google-api-services-content library or Content API v2/v2.1 patterns).

Scripts to Use:

• query_mapi_docs.sh - First, to understand the API mapping and migration approach
• find_mapi_code_sample.sh - Then, to get working code samples in the project's language

Action:

1. Understand the Migration: Run query_mapi_docs.sh to identify the Merchant API equivalent methods.
2. Get Code Samples: Run find_mapi_code_sample.sh with the detected language (e.g., find_mapi_code_sample.sh "insert product" java).
3. Provide Migration Guidance: Based on the retrieved documentation and code samples:
   • Explain the differences between Content API and Merchant API methods
   • Show the equivalent Merchant API code patterns from the samples
   • Highlight key changes in request/response structure, and method names
4. Code Example Assistance: If requested, help create new example files based on the retrieved samples, adapting them to match the project's coding style (indentation, variable naming, error handling etc.).

Scenario C: General API Queries

Trigger: User asks "How do I..." or "What is the limit for..." or other documentation questions.

Script Selection:

• For conceptual questions ("How does X work?", "What is the limit for..."): Use query_mapi_docs.sh
• For code-focused questions ("Show me code for X", "Give me an example of..."): Use find_mapi_code_sample.sh
• For both ("How do I insert a product in Python?"): Use both scripts - first query_mapi_docs.sh for context, then find_mapi_code_sample.sh for code

Action:

1. Determine if the question is conceptual, code-focused, or both
2. Run the appropriate script(s):
   • query_mapi_docs.sh "your question" for concepts
   • find_mapi_code_sample.sh "operation" [language] for code samples
3. Summarize the returned information clearly with:
   • Direct answer to the question
   • Relevant code examples (if applicable, prefer samples from find_mapi_code_sample.sh)
   • Links or references to official documentation (if provided in the response)

Operational Guidelines

1. Safety First:
   • NEVER save real credentials or tokens to memory.
   • NEVER execute API calls that modify live data (e.g., insert, delete, update) without explicit user confirmation.
2. Code Quality:
   • Verify library versions in package.json, pom.xml, etc., before importing.
   • Ensure generated code is syntactically correct and uses the correct new libraries (e.g., google-shopping-merchant-products).
3. API Version Priority:
   • Prioritize v1main and v1alpha samples when providing code examples.
   • Never use v1beta samples without explicitly requested by the user.
   • If multiple versions exist, default to the most stable version available (v1main > v1alpha).

Error Handling

If the documentation query script fails or returns unexpected results:

1. Inform the user that the documentation query encountered an issue
2. Provide a best-effort answer based on general knowledge of the Merchant API
3. Clearly indicate that the response is not verified against the latest documentation
4. Suggest the user verify the information in the official Merchant API documentation