Discovers and integrates third-party APIs using the context-matic MCP server. Uses `fetch_api` to find available API SDKs, `ask` for integration guidance, `model_search` and `endpoint_search` for SDK details. Use when the user asks to integrate a third-party API, add an API client, implement features with an external API, or work with any third-party API or SDK.
When the user asks to integrate a third-party API or implement anything involving an external API or SDK, follow this workflow. Do not rely on your own knowledge for available APIs or their capabilities — always use the context-matic MCP server.
Apply this skill when the user:
Before checking for guidelines or skills, identify the project's primary programming language by inspecting the workspace:
| File / Pattern | Language |
|---|---|
| *.csproj, *.sln | csharp |
| package.json with "typescript" dep or .ts files | typescript |
| requirements.txt, pyproject.toml, *.py | python |
| go.mod, *.go | go |
| pom.xml, build.gradle, *.java | java |
| Gemfile, *.rb | ruby |
| composer.json, *.php | php |
Use the detected language in all subsequent steps wherever language is required.
Check whether guidelines and skills have already been added for this project by looking for their presence in the workspace.
{language}-conventions is the skill produced by add_skills.{language}-security-guidelines.md and {language}-test-guidelines.md are language-specific guideline files produced by add_guidelines.update-activity-workflow.md is a workflow guideline file produced by add_guidelines (it is not language-specific).{language}-conventions is missing for the project's language: Call add_skills.{language}-conventions already exist: Skip this step and proceed to step 2.Call fetch_api to find available APIs — always start here.
language parameter using the language detected in step 1a.key parameter: pass the API name/key from the user's request (e.g. "paypal", "twilio").fetch_api with that value.key) when there is no exact match.key for the user's requested API before proceeding. This key will be used for all subsequent tool calls related to that API.If the requested API is not in the list:
ask with: language, key (from step 2), and your query.These tools return definitions only — they do not call APIs or generate code.
language, key, and an exact or partial case-sensitive model name as query (e.g. availableBalance, TransactionId).language, key, and an exact or partial case-sensitive method name as query (e.g. createUser, get_account_balance).Call update_activity (with the appropriate milestone) whenever one of these is concretely reached in code or infrastructure — not merely mentioned or planned:
| Milestone | When to pass it |
|---|---|
| sdk_setup | SDK package is installed in the project (e.g. npm install, pip install, go get has run and succeeded). |
| auth_configured | API credentials are explicitly written into the project's runtime environment (e.g. present in a .env file, secrets manager, or config file) and referenced in actual code. |
| first_call_made | First API call code written and executed |
| error_encountered | Developer reports a bug, error response, or failing call |
| error_resolved | Fix applied and API call confirmed working |
add_guidelines called if guideline files were missing, otherwise skippedadd_skills called if {language}-conventions was missing, otherwise skippedfetch_api called with correct language and key (API name)key identified for the requested API (or user informed if not found)update_activity called only when a milestone is concretely reached in code/infrastructure — never for questions, searches, or tool lookupsupdate_activity called with the appropriate milestone at each integration milestoneask used for integration guidance and code samplesmodel_search / endpoint_search used as needed for SDK detailsfetch_api, do not guess at SDK usage — inform the user that the API is not currently available in this plugin and stop.fetch_api is API discovery, not integration — do not call update_activity before it.Copy a source-pinned command for your client. You run it yourself.
Destination: .claude/skills/integrate-context-matic · pinned to the source commit
# Run from your project root
git clone https://github.com/github/awesome-copilot.git .skillboard-tmp
git -C .skillboard-tmp checkout f11a4e441c5ff061b4f8ae37952be8c602e4034e
mkdir -p ".claude/skills"
cp -r ".skillboard-tmp/skills/integrate-context-matic" ".claude/skills/"
rm -rf .skillboard-tmpReview the source before running. This copies files into your project; it is not a one-click install and does not verify runtime safety.
sudo apt update && sudo apt install -y gitnpm install -g @anthropic-ai/claude-code# Run from your project root
git clone https://github.com/github/awesome-copilot.git .skillboard-tmp
git -C .skillboard-tmp checkout f11a4e441c5ff061b4f8ae37952be8c602e4034e
mkdir -p ".claude/skills"
cp -r ".skillboard-tmp/skills/integrate-context-matic" ".claude/skills/"
rm -rf .skillboard-tmpDestination: .claude/skills/integrate-context-matic
Scanner static-checks@0.1.0 · commit f11a4e441c5f. Static checks cannot prove runtime safety – review the source and the exact diff before installing. How checks work.
References credentials, tokens or secret files that a skill should not need.
Evidence: [redacted]· fingerprint e9cbb0224c4a3d23
Instructs shell/process/package operations that run commands on the host.
Evidence: npm install· fingerprint 3a2dc0ae21eb56d7