tools-and-agent

Rules describing how an agent system coordinates messages, shell commands, web searches, history, cancellation, and repeated tool calls. They also define the common success-and-error format returned by tools.

In plain words
What is it for?
Use them when implementing or reviewing the agent orchestrator, shell and search tools, streaming responses, tool-call rounds, cancellation, and consistent tool-result handling.
Why use it?
They give developers a shared contract for how the agent loop and its tools should behave, including how errors are reported and when processing stops.

Cursor rule for Cursor

Install

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.

agentmods
npx agentmods add rules/superagent-ai/grok-cli/tools-and-agent
Clone the repo
git clone --depth 1 https://github.com/superagent-ai/grok-cli

Made for: Cursor.

Per session 0 Nothing until a file matches its globs; then the whole rule loads.
When invoked 475 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What 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.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.00475
Opus 5 $0.00000 $0.00237
Sonnet 5 $0.00000 $0.00095
Haiku 4.5 $0.00000 $0.00047

Measured 2d ago against content hash 347d4f1d026e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

tools-and-agent scanned grade A with 0 findings 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.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.cursor/rules/tools-and-agent.mdc · 63 lines

What it actually says

Agent & Tools System

Agent

Agent (in src/agent/agent.ts) is the core orchestrator. It manages:

  • The Grok API client (GrokClient)
  • Bash tool for all shell operations
  • Web search and X search via the Responses API
  • Chat history and message accumulation
  • Abort/cancellation support

Agent Loop

The agent uses an iterative tool-call pattern via processMessage (async generator):

  1. Send messages to the LLM via Chat Completions API (streaming).
  2. If the response contains tool_calls, execute each tool.
  3. Append tool results to the message history.
  4. Repeat until no tool calls remain or maxToolRounds is reached.
  5. Yield StreamChunk objects for real-time UI updates.

Tool Interface

All tools return a ToolResult:

interface ToolResult {
  success: boolean;
  output?: string;
  error?: string;
}

Tools should never throw — wrap errors in { success: false, error: "..." }.

Built-in Tools

Tool File Purpose
BashTool src/tools/bash.ts Execute any shell command
search_web via GrokClient.searchWeb() Web search using Responses API
search_x via GrokClient.searchX() X/Twitter search using Responses API

Grok Client

GrokClient (in src/grok/client.ts) handles two API endpoints:

  • Chat Completions (/v1/chat/completions): Main agent loop with streaming and tool calling
  • Responses API (/v1/responses): X Search and Web Search with server-side tools

Adding a New Tool

  1. Add the tool schema to src/grok/tools.ts (in the TOOLS array).
  2. Add the execution case in Agent.executeTool().
  3. Update the system prompt in Agent to document the tool.
Changes

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.

  1. 2d ago First seen · 63 lines · 0 tokens per session scan A 347d4f1d026e

Subscribe to this mod's changes

tools-and-agent is a cursor rule published in the GitHub repository superagent-ai/grok-cli (3,445 stars, last pushed 1mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 475 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.