AAIF Public Agents

This repository contains public, reusable agent resources for AAIF, including skills, recipes, prompts, and agent-facing workflows.

Repository Structure

public-agents/
├── skills/
│   └── ...
├── recipes/
│   └── ...
├── prompts/
│   └── ...
└── docs/
    └── ...

Contents

• skills/ — Agent Skills for reusable domain expertise, brand guidance, workflows, and tool-specific instructions.
• recipes/ — Repeatable agent workflows and task definitions.
• prompts/ — Reusable prompts and prompt templates.
• docs/ — Supporting documentation for contributors and users.

Usage

Start with the relevant directory for what you need:

• Looking for installable agent capabilities? See skills/.
• Looking for repeatable workflows? See recipes/.
• Looking for reusable prompt templates? See prompts/.

GitHub Actions Workflows

Manually-triggered workflows in .github/workflows/ that call LLM APIs.

copilot-test-action

Calls the GitHub Models inference API (https://models.github.ai) using the built-in GITHUB_TOKEN. Lists all available model IDs from the catalog, then sends a "hello world" prompt to whichever model you pick.

Requirements:

• The org must have GitHub Models enabled in Org Settings → Models. Without this, inference returns 403 even though the catalog endpoint works.
• Workflow uses the org-scoped endpoint (/orgs/{org}/inference/chat/completions).

No secret setup required — GITHUB_TOKEN is provided automatically.

gemini-test-action

Calls the Google Gemini API directly with your own API key. Sends a "hello world" prompt to a configurable Gemini model and prints the response.

Setup: storing the Gemini API key

1. Get a key from Google AI Studio.
2. Add it as a repository secret named GEMINI_API_KEY:
   • Go to Repo → Settings → Secrets and variables → Actions → New repository secret
   • Name: GEMINI_API_KEY
   • Secret: (paste the key)
   • Click Add secret

That's it. The workflow reads it from ${{ secrets.GEMINI_API_KEY }} and passes it via the x-goog-api-key header (never in the URL, so it won't appear in logs or referrer headers).

Running it

1. Go to Actions → gemini-test-action → Run workflow.
2. Pick a model (default gemini-2.5-flash; other options include gemini-2.5-pro, gemini-2.0-flash, gemini-1.5-pro, gemini-1.5-flash).
3. Click Run workflow.

The log will show the HTTP status, raw JSON response, and the extracted text output.

Scoping the secret

For tighter blast radius, you can store GEMINI_API_KEY in a GitHub Environment instead of repo-wide secrets, then add environment: <name> to the job. Useful if you later want approval gates or want different keys for different branches.