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 rules/iloveitaly/llm-ide-rules/generalgit clone --depth 1 https://github.com/iloveitaly/llm-ide-rulesWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/rules/iloveitaly/llm-ide-rules/general)<a href="https://agentmods.dev/rules/iloveitaly/llm-ide-rules/general"><img src="https://agentmods.dev/badge/rules/iloveitaly/llm-ide-rules/general.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00992 | $0.00992 |
| Opus 5 | $0.00496 | $0.00496 |
| Sonnet 5 | $0.00198 | $0.00198 |
| Haiku 4.5 | $0.00099 | $0.00099 |
Grade A, and why
general 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 5d 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.
How it starts
The opening of the file, as written. The whole thing — 72 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Coding instructions for all programming languages:
- Never use emojis anywhere unless explicitly requested.
- If no language is specified, assume the latest version of python.
- If tokens or other secrets are needed, pull them from an environment variable
- Prefer early returns over nested if statements.
- Prefer
continuewithin a loop vs nested if statements. - Prefer smaller functions over larger functions. Break up logic into smaller chunks with well-named functions.
- Prefer constants with separators:
10_000is preferred to10000(or10_00over1000in the case of a integer representing cents). - When I ask you to write code, prioritize simplicity and legibility over covering all edge cases, handling all errors, etc.
- When a particular need can be met with a mature, reasonably adopted and maintained package, I would prefer to use that package rather than engineering my own solution.
- Never add error handling to catch an error without being asked to do so. Fail hard and early with assertions and allow exceptions to propagate.
- When naming variables or functions, use names that describe the effect. For example, instead of
function handleClaimFreeTicket(a function which opens a dialog box) usefunction openClaimFreeTicketDialog. - Do not install missing system packages! Instead, ask me to install them for you.
- If terminal commands are failing because of missing variables or commands which are unrelated to your current task, stop your work and let me know.
- Don't worry about fixing lint errors or running lint scripts unless I specifically ask you to.
- When implementing workarounds for tooling limitations (like using
Anyfor unresolvable types) or handling non-obvious edge cases, always add a brief inline comment explaining the technical reasoning. - Reserve exact-width
#-box section separators for long files requiring organization, though they should not be necessary in the large majority of cases (separate files is generally better).
Use line breaks to organize code into logical groups. Instead of:
if not client_secret_id:
raise HTTPException(status.HTTP_400_BAD_REQUEST)
session_id = client_secret_id.split("_secret")[0]
Prefer:
if not client_secret_id:
raise HTTPException(status.HTTP_400_BAD_REQUEST)
session_id = client_secret_id.split("_secret")[0]
DO NOT FORGET: keep your responses short, dense, and without fluff. I am a senior, well-educated software engineer, and hate long explanations.
Add Comments for Expert Engineer with Limited Domain Knowledge
The engineer reading your code is a world-class software engineer, but is not familiar with the internals of every system. Include concise one-line comments explaining key hooks, API usage, blocks of logic, etc., to help the reader quickly understand the code you've written.
In other words, embed the business requirements as comments in the code when the code does not self-document.
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.
- 5d ago First seen · 72 lines · 992 tokens per session scan A ef52282cfa3b
general is a cursor rule published in the GitHub repository iloveitaly/llm-ide-rules (13 stars, last pushed 4d ago), licensed MIT. It adds 992 tokens to every session, about $0.0050 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 2026-08-30.
Other cursor rules, from other repositories
ai-kb
Use shared AI KB rules and context.
read-xlsx
Reading, writing, diffing, and repairing spreadsheets (.xlsx) for AI agents via the xfa MCP server.
codesynapse-cli
Use this rule to answer codebase architecture questions, trace dependencies, find callers/callees, calculate blast radius, and explore the code graph. Triggers on: "how does X work", "what handles Y", "where is Z defined", "explain the mechanism", "who calls", "blast radius", "shortest path", "trace request flow"…
cdv
How to drive the cdv MCP sidecar (observe, verify, remember, resume).
pn-tool-risk-policy
MCP tool risk tiers, HITL gates, and when to use approvalcheckpoint and gatelogappend. Use for safety reviews and workflow audits.
004_auto_context
Rules for requesting relevant memory context based on file types, locations, and current activity.