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.
git clone --depth 1 https://github.com/tobitege/Flowery.UnoWrote 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/tobitege/flowery.uno/must-read)<a href="https://agentmods.dev/rules/tobitege/flowery.uno/must-read"><img src="https://agentmods.dev/badge/rules/tobitege/flowery.uno/must-read/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/rules/tobitege/flowery.uno/must-read"><img src="https://agentmods.dev/badge/rules/tobitege/flowery.uno/must-read.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00501 | $0.00501 |
| Opus 5.5 | $0.00200 | $0.00200 |
| Sonnet 5.5 | $0.00100 | $0.00100 |
| Haiku 4.5 | $0.00050 | $0.00050 |
Grade A, and why
must-read 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 13d 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 — 61 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Must Read
⛔ 3-Iteration Rule (No Silent Scope/Approach Changes)
During multi-tool execution: If an approach isn't working after 3 attempts, ASK before trying something different. Do NOT:
- Silently decide "this is too complex, let me rewrite/simplify it"
- Expand scope from a targeted fix to a multi-hundred-line refactor
- Switch fundamental approaches mid-execution without user awareness
The rule prevents the AI from unilaterally changing direction. Interactive discussion with the user is always fine.
"Fix Root Cause" vs "Minimal Change" — These are NOT contradictory:
- ✅ Fix root causes across multiple files if needed — just mention what you're doing
- ❌ Don't refactor unrelated code "while you're there"
- ⚠️ Pause and ask only for significant scope changes (10-line fix → 200-line rewrite, architecture changes)
⚠️ Git Bash CRLF Rule
When using sed, cat, head, tail on Windows files, ALWAYS pipe through | tr -d '\r':
sed -n '100,110p' file.cs | tr -d '\r'
Otherwise output will be garbled due to CRLF line endings.
"No Change" Is Valid
When investigating issues without a specified fix, ASK before coding. Don't go on a "coding spree" - the behavior may be intentional.
Search Tool Quirks
Some agents' search tools fail silently on single-file paths. If search returns 0 results unexpectedly:
- Try searching a directory instead, with a filter for the specific file
- Fall back to terminal
rgif the native tool misbehaves
Terminal Search
When native search tools fail or for complex regex searches, use terminal rg (ripgrep):
clear && rg "pattern" path/to/search
Code Generation
- No Trailing Whitespace: Never generate lines with trailing spaces or tabs
Refactoring Standards
- Property/Field Sync: When renaming/deleting, search file-wide for ALL references
- Verify Infrastructure: Check constructors/fields exist before using them
- Constructor Fallback: Use property initializers
{ Prop = value }if no matching constructor
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.
- 13d ago First seen · 61 lines · 501 tokens per session scan A f1f26a1a5697
must-read is a cursor rule published in the GitHub repository tobitege/Flowery.Uno (16 stars, last pushed today), licensed MIT. It adds 501 tokens to every session, about $0.0020 per session on Opus 5.5. 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 2026-10-02.
Other cursor rules, from other repositories
roslyn-mcp-overview
RoslynMcpServer — architecture, layout, version, DI.
roslyn-mcp-pitfalls
Known pitfalls and regression lessons from production use.
roslyn-mcp-docs-agents
Docs sync AGENTS/README, agent conventions, PowerShell, secrets.
roslyn-mcp-tools
MCP tool mapping — skeleton vs decompile vs symbols vs build.
roslyn-mcp-version-bump
Auto-bump Version after successful MCP self-build and reload.
roslyn-mcp-decompile
Assembly resolve, ILSpy decompile, NuGet/BCL fallback.