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/girijashankarj/cursor-handbook/token-efficiencygit clone --depth 1 https://github.com/girijashankarj/cursor-handbookWrote 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/girijashankarj/cursor-handbook/token-efficiency)<a href="https://agentmods.dev/rules/girijashankarj/cursor-handbook/token-efficiency"><img src="https://agentmods.dev/badge/rules/girijashankarj/cursor-handbook/token-efficiency.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 | $0.00540 | $0.00540 |
| Opus 5 | $0.00270 | $0.00270 |
| Sonnet 5 | $0.00108 | $0.00108 |
| Haiku 4.5 | $0.00054 | $0.00054 |
Grade A, and why
token-efficiency 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 today.
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 — 52 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Token Efficiency Rules
Context Layering Strategy
Use layered context to minimize token consumption:
| Layer | Token Budget | What to Include |
|---|---|---|
| Immediate | {{CONFIG.cursor.tokenOptimization.contextLayers.immediate}} | Current file, direct dependencies |
| Relevant | {{CONFIG.cursor.tokenOptimization.contextLayers.relevant}} | Related modules, shared types |
| Extended | {{CONFIG.cursor.tokenOptimization.contextLayers.extended}} | Broader codebase patterns |
| Reference | {{CONFIG.cursor.tokenOptimization.contextLayers.reference}} | Documentation, examples |
Expensive Operations — REQUIRE CONFIRMATION
Operations that MUST require user confirmation:
- Full test suite (
{{CONFIG.testing.testCommand}}) — estimated 100K+ tokens - Full lint (
{{CONFIG.testing.lintCommand}}) — estimated 50K+ tokens - Full build — estimated 75K+ tokens
- Reading 10+ files — estimated 50K+ tokens
Preferred alternatives:
- Use
read_lintstool instead of running lint commands - Run
{{CONFIG.testing.typeCheckCommand}}(~10K tokens) instead of full tests - Run tests for a single file instead of the entire suite
- Read only the specific functions/sections needed, not entire files
Output Guidelines
- Summaries over verbose: Use bullet points, not paragraphs
- Diff format: Show only changed lines when editing files
- Max 3 sentences for explanations unless user asks for more
- No redundant context: Don't repeat information the user already provided
- Structured responses: Use headers, lists, and tables
- Code blocks: Only show relevant sections, not entire files
File Reading Strategy
- Read only the files directly related to the task
- Start with type definitions and interfaces
- Read implementation files only when needed
- Never read generated files (dist/, build/, coverage/)
- Use file search tools to find specific symbols instead of reading entire files
Token-Saving Patterns
- When asked to "review" code, focus on: bugs, security issues, performance — skip style
- When asked to "implement", show the minimal diff — not the entire file
- When asked to "explain", use a layered approach: summary first, details on request
- When asked to "test", create tests for the specific function — not the entire module
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.
- today First seen · 52 lines · 540 tokens per session scan A 0f37957a7a71
token-efficiency is a cursor rule published in the GitHub repository girijashankarj/cursor-handbook (30 stars, last pushed 3d ago), licensed MIT. It adds 540 tokens to every session, about $0.0027 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 2026-09-03.
Other cursor rules, from other repositories
graphstack
GraphStack v4.7.1 — Orchestrated, graph-first, GNAP-tracked AI development (cross-platform).
new_features
Guidelines for integrating new features into the Task Master CLI.
utilities
// ✅ DO: Create focused, reusable utilities /.
tasks
Guidelines for implementing task management operations.
telemetry
Guidelines for integrating AI usage telemetry across Task Master.
context_gathering
Standardized patterns for gathering and processing context from multiple sources in Task Master commands, particularly for AI-powered features.