cursorrules

A set of coding rules for a speech-focused Python and React project. It covers voice services, code style, tool responses, testing, and how the agent should use speech documentation.

In plain words
What is it for?
Use it when changing the speech MCP: checking service availability, finding speech-related documentation, running lint and tests, and recording useful findings.
Why use it?
It gives the coding agent consistent project conventions and checks to follow, reducing avoidable style, integration, and testing mistakes.

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/sandraschi/speech-mcp/cursorrules
Clone the repo
git clone --depth 1 https://github.com/sandraschi/speech-mcp

Made for: Cursor.

Per session 255 This file is loaded in full into every session.
When invoked 255 The same file — it is already loaded in full.
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 $0.00255 $0.00255
Opus 5 $0.00128 $0.00128
Sonnet 5 $0.00051 $0.00051
Haiku 4.5 $0.00026 $0.00026

Measured yesterday against content hash 2c43b3be0cf9, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

cursorrules 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 yesterday.

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.

.cursorrules · 22 lines

What it actually says

Cursor Rules — speech-mcp

  • Python: uv run ruff check src/ --fix before commit; line length 120.
  • Type hints: Annotated[T, Field(description="...")] — no Args: blocks.
  • Tools return {"success": bool, ...} dicts; list/status tools use Prefab UI (@mcp.tool(app=True)).
  • Never call os.getenv() outside server.py; never block the event loop.
  • Frontend: React + Vite + Tailwind (dark theme only), Biome for lint/format.
  • Tests: uv run pytest tests/ -m "not live".

Session Context (Speech MCP)

You have access to the fleet voice gateway: TTS (windows/gemini/hume/elevenlabs/gemma), local STT (FunASR), offline wake word, RAG over speech docs, and a Voice Command Bus into fleet-agent.

Before starting work:

  1. Check provider availability: fleet_health_overview()
  2. Find speech docs: search_docs(query="FunASR or wake word")

At end of work, save insights:

  • Ingest useful speech findings into the RAG knowledge base
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. yesterday First seen · 22 lines · 255 tokens per session scan A 2c43b3be0cf9

Subscribe to this mod's changes

cursorrules is a cursor rule published in the GitHub repository sandraschi/speech-mcp (2 stars, last pushed 13d ago), licensed MIT. It adds 255 tokens to every session, about $0.0013 per session on Opus 5. 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 2026-08-31.