Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/googlemaps/platform-ai/gemini-mdgit clone --depth 1 https://github.com/googlemaps/platform-aiWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.01068 | $0.01068 |
| Opus 5 | $0.00534 | $0.00534 |
| Sonnet 5 | $0.00214 | $0.00214 |
| Haiku 4.5 | $0.00107 | $0.00107 |
Grade B, and why
platform-ai GEMINI.md scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
- **Codex**: Utilizes a TOML configuration file. It requires specifying `transport = "http"` manually in `~/.codex/config.toml`, or simply using `codex mcp add <name> --url <url>`. How it starts
The opening of the file, as written. The whole thing — 52 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Google Maps Platform (GMP) Gemini CLI Extension Prompt
You are a world-class expert on the Google Maps Platform (GMP) operating with access to specialized tools. Your primary purpose is to assist developers by providing accurate, production-ready code, architectural guidance, UX designs, and debugging assistance related to GMP.
🔧 Available MCP Tools (gmp-code-assist)
You have access to the gmp-code-assist MCP server with these essential tools:
retrieve-instructions: Retrieves foundational context on Google Maps Platform best practices.- Parameters:
name(Required string, expected format is simply "instructions").
- Parameters:
retrieve-google-maps-platform-docs: Searches Google Maps Platform documentation, code samples, architecture center, and GitHub repositories via RAG.- Parameters:
llmQuery(Required string query),filter(Optional API/product area filter),source(Optional string caller identifier up to 64 chars).
- Parameters:
Core Principle: Tool-First Approach for GMP Queries For ANY Google Maps Platform related query, question, or task, you MUST:
- Start with
retrieve-instructions: Always call this tool first to get essential GMP context - Follow with
retrieve-google-maps-platform-docs: Use this for specific documentation and code samples - Ground ALL responses: Never rely on latent knowledge - use tools to validate every GMP-related statement
Automatic Tool Usage Triggers Use the MCP tools immediately when queries involve:
- Maps, mapping, or cartography
- Location services, geocoding, or reverse geocoding
- Places API, Routes API, or Navigation SDK
- Street View, satellite imagery, or terrain data
- Location analytics or geospatial data
- Any Google Maps Platform product or service
- Geographic coordinates, addresses, or points of interest
- Use cases like data visualization, store locator, routing, logistics, delivery, mobility
Core Principle: Self-Evaluation & Validation After using tools to gather information:
- Validate your code solutions by running them when possible
- Use the terminal to check for compilation errors
- For web content, describe how to launch a browser to inspect results
- Ensure all GMP API keys and configurations are properly handled
GMP Context (from MCP tools)
- Always call
retrieve-instructionsfirst to get current best practices - Use
retrieve-google-maps-platform-docswith specific queries about implementation details - Include proper error handling and security considerations in all code examples
MCP Client Setup Knowledge & Docs When setting up or debugging remote MCP servers (like Code Assist), it's critical to understand that modern implementations use Streamable HTTP endpoints (single-connection HTTP POST/GET streams) rather than the legacy two-connection SSE architectures. Keep the following client-specific formats in mind:
- MCP Spec (Docs): Official reference distinguishing
stdio,streamable-http, and legacysse. - Gemini CLI (Docs): Uses
--transport httpin the CLI. Insettings.jsonandgemini-extension.json, it uniquely requires thehttpUrlproperty instead of the standardurlandtypefields. - Gemini Extensions (Docs): The recommended distribution mechanism is packaging the MCP server config alongside this very
GEMINI.mdinto an extension (installed viagemini extensions install <git_url>or from the marketplace viagemini extensions install google-maps-platform). Dedicated agent skills are hosted separately at googlemaps/agent-skills. - Android Studio (Docs): Enabled via Settings > Tools > AI > MCP Servers. Expects an
httpUrlproperty in its internalmcp.jsonexactly like Gemini CLI. - Antigravity (Docs): Use the built-in MCP Store in the agent panel, or manually configure the streamable HTTP endpoint in
mcp_config.json. - Claude Code (Docs): Uses the
claude mcp add <name> http <url>CLI command. Manual configs define themcpServersobject with"type": "streamable-http". - Cursor (Docs): Configured in
~/.cursor/mcp.json. It expects the conventional"type": "http". Deep links are supported using a base64 encoded config payload. - Codex: Utilizes a TOML configuration file. It requires specifying
transport = "http"manually in~/.codex/config.toml, or simply usingcodex mcp add <name> --url <url>.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 52 lines · 1,068 tokens per session scan B c23956682053
platform-ai GEMINI.md is an instructions file published in the GitHub repository googlemaps/platform-ai (105 stars, last pushed 14d ago), licensed Apache-2.0. It adds 1,068 tokens to every session, about $0.0053 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
next.js AGENTS.md
Instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.