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/ghaida/intent/intent-ref-interaction-patternsgit clone --depth 1 https://github.com/ghaida/intentWhat 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 | $0.00039 | $0.03271 |
| Opus 5 | $0.00019 | $0.01636 |
| Sonnet 5 | $0.00008 | $0.00654 |
| Haiku 4.5 | $0.00004 | $0.00327 |
Grade A, and why
intent-ref-interaction-patterns 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.
How it starts
The opening of the file, as written. The whole thing — 236 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Interaction Patterns
Form Design Principles
Forms are where users exchange value with your product. Every unnecessary field, confusing label, or unhelpful error message is friction between the user and their goal. The research on form design is extensive and remarkably consistent.
One Thing Per Page
The Government Digital Service (GDS) pattern, validated across millions of transactions: each screen asks one question or collects one piece of information. Not one form field — one conceptual unit.
Why it works: Reduces cognitive load. Each page has a clear purpose. Error recovery is simpler — the error is on this page, about this thing. Progress feels tangible. Mobile performance improves (less content per load). Analytics are more granular (you know exactly where drop-off occurs).
When to bend it: Closely related fields that users think of as one concept (first name + last name, city + state + zip). Editing flows where users need to see multiple fields to understand context. Expert tools where speed matters more than guidance.
When not to bend it: Checkout flows. Registration. Any flow where drop-off is a risk. Any flow used on mobile.
Inline Validation Timing
When to validate is as important as how to validate. Get the timing wrong and validation becomes harassment.
Validate on blur (leaving a field), not on keystroke. Validating while the user is still typing is hostile — they haven't finished their input and you're already telling them it's wrong. Luke Wroblewski's research (2009) confirmed that inline validation improves completion rates, but only when triggered after the user leaves the field.
Exception: password strength. Real-time feedback on password requirements is one of the few cases where keystroke-level validation helps, because the user is building toward a goal and needs to know the criteria as they type.
Exception: character counts. If a field has a maximum length, show the remaining count as the user types. Don't wait until they've written a paragraph to tell them the limit is 140 characters.
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 · 236 lines · 3,271 tokens per session scan A 084e599696ab
intent-ref-interaction-patterns is a cursor rule published in the GitHub repository ghaida/intent (139 stars, last pushed 1mo ago), licensed CC0-1.0. It adds 39 tokens to every session and 3,271 once invoked, about $0.0002 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
angular-20
This rule provides comprehensive best practices and coding standards for Angular development, focusing on modern TypeScript, standalone components, signals, and performance optimizations.
dev-standard
Apache Superset development standards and guidelines for Cursor IDE.
typescript
Changes to these high-fan-out internals can affect every message, delta, element, or rerun. Keep work in them minimal, and benchmark changes with representative stress-test apps.
coolify-ai-docs
Master reference to all Coolify AI documentation in .ai/ directory.
python_lib
Tips and guidelines specific to the development of the Streamlit Python library, not applicable to scripts and e2e tests.
specs
This directory contains product and tech specs for Streamlit features.