copilot-conventions

A set of coding rules for AI assistants working with code, including reading before editing, keeping changes focused, testing, and following Git practices.

In plain words
What is it for?
Use it to guide coding, debugging, reviewing, and testing work across tools such as GitHub Copilot, Claude Code, Cursor, or local language models.
Why use it?
It helps prevent unnecessary changes, guesses, unverified bug fixes, and code that does not match the project's existing structure.

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/gosha70/code-copilot-team/copilot-conventions
Clone the repo
git clone --depth 1 https://github.com/gosha70/code-copilot-team

Made for: Cursor.

Per session 1,259 This file is loaded in full into every session.
When invoked 1,259 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 $0.01259 $0.01259
Opus 5 $0.00629 $0.00629
Sonnet 5 $0.00252 $0.00252
Haiku 4.5 $0.00126 $0.00126

Measured 2d ago against content hash c123ccc17621, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

copilot-conventions 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.

adapters/cursor/.cursor/rules/copilot-conventions.mdc · 89 lines

How it starts

The opening of the file, as written. The whole thing — 89 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Cross-Copilot Conventions

Shared rules that apply whether using Claude Code, GitHub Copilot, Cursor, or local LLMs. These conventions ensure consistent behaviour regardless of which AI tool is driving.

Core Contract

  1. Read before write — understand existing code and patterns first.
  2. Minimal changes — only modify what was requested.
  3. Show your work — explain changes, provide diffs.
  4. Test everything — run linters and tests after code changes.
  5. Ask when uncertain — do not guess at ambiguous requirements.
  6. Verify before diagnosing — when asked to fix a reported bug, re-run the failing test or reproduce the symptom first. The issue may already be fixed. Do not spend time diagnosing a problem that no longer exists.

Target Alignment and Complexity Control

Before planning, implementing, or approving agent-produced work:

  1. Restate the originating goal and concrete acceptance criteria. Derived plans and agent summaries do not replace the original request.
  2. Map every changed file and each new abstraction, interface, state, dependency, or test layer to a current requirement or a concrete correctness or safety failure. Remove or defer anything without that link.
  3. Prefer the smallest solution that fits the existing architecture. Do not turn narrow work into a framework, generalized engine, speculative state model, broad hardening exercise, or unrelated cleanup.
  4. Keep tests, documentation, and proof proportional to the risk and acceptance criteria. More artifacts are not evidence of better alignment.
  5. Complete the required path before optional improvements. Record useful extras as follow-up work instead of silently expanding scope.
  6. Before approval, ask: What can be removed without weakening correctness, safety, or the acceptance criteria?

For a handoff from Claude Code or any other agent, explicitly classify the change as on-target, overcomplicated, or off-target, and explain the evidence. Green tests do not compensate for scope drift. Judge the artifact rather than guessing at a vendor's intent; complexity required by a concrete failure mode is not overengineering when the link is explicit.

Read the full file on GitHub · 89 lines

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 · 89 lines · 1,259 tokens per session scan A c123ccc17621

Subscribe to this mod's changes

copilot-conventions is a cursor rule published in the GitHub repository gosha70/code-copilot-team (6 stars, last pushed 2d ago), licensed MIT. It adds 1,259 tokens to every session, about $0.0063 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.