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/Intrafere/MOTO-Autonomous-ASIWrote 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/intrafere/moto-autonomous-asi/rag-design-for-overall-program)<a href="https://agentmods.dev/rules/intrafere/moto-autonomous-asi/rag-design-for-overall-program"><img src="https://agentmods.dev/badge/rules/intrafere/moto-autonomous-asi/rag-design-for-overall-program/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/intrafere/moto-autonomous-asi/rag-design-for-overall-program"><img src="https://agentmods.dev/badge/rules/intrafere/moto-autonomous-asi/rag-design-for-overall-program.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 | $0.00019 | $0.05746 |
| Opus 5 | $0.00010 | $0.02873 |
| Sonnet 5 | $0.00004 | $0.01149 |
| Haiku 4.5 | $0.00002 | $0.00575 |
Grade A, and why
rag-design-for-overall-program 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 9d 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 — 268 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Important Notes When Editing RAG Systems
The RAG system in this program is very advanced, be certain that any changes you make to the RAG system are correct changes.
Canonical Direct-Injection / RAG Policy
Direct injection is preferred for source content, but the allocator may offload an otherwise-fitting optional block to RAG when direct injection would starve the reserved evidence budget. Code currently keeps a 5000-token reserve for RAG/evidence in the shared context allocator.
Some inputs are mandatory direct-inject and must never be RAG'd, summarized, compressed, truncated, excerpted, or replaced by partial views. If mandatory direct-inject context does not fit the configured model context, halt with an explicit context-overflow error and tell the user which mandatory context overflowed.
If an item is direct injected, its RAG counterpart must NOT also be included.
The active Progressive Solution Path is never RAG-indexed. Before five accepted brainstorm ideas it is absent; afterward, only the bounded Main Submitter 1-approved canonical revision is optional advisory direct context for relevant solving and semantic-validator calls. It may be shed before mandatory user/source/candidate/tool context when budget requires, and pending, rejected, stale, and historical revisions never enter model context. The path is not evidence and may always be ignored for a better route.
Paper-Writing / Research Offload Order
These priorities apply to Aggregator, Compiler, and Autonomous paper-writing workflows. They do not describe LeanOJ proof-only memory ordering.
| Mode | Mandatory direct context | Optional direct-first blocks | RAG/offload order |
|---|---|---|---|
| Aggregator submitter | User prompt, role/schema instructions | Shared training DB, local submitter DB, rejection log, user uploads | Shared Training DB → Local Submitter DB → Rejection Log → User Upload Files |
| Aggregator validator | User prompt, role/schema instructions, submission(s) under review | Shared training DB, user uploads | Shared Training DB → User Upload Files |
| Aggregator cleanup review | User prompt, cleanup schema/instructions | Accepted-submissions DB if it fits with reserve | Accepted-submissions DB via normal RAG fallback; cleanup never skips solely because DB is large |
| Compiler construction/review | User prompt, current outline | Current paper, autonomous brainstorm/source DB when injected by the caller, rejection/acceptance logs | Reference Papers → Brainstorm/Source DB → Current Paper → Rejection/Acceptance Logs |
| Compiler outline update | User prompt, current outline | Current paper, rejection/acceptance logs | Current Paper → Rejection/Acceptance Logs |
| Compiler rigor | User prompt, current outline, current paper, Lean candidate/attempt context | Existing verified-proof summaries and recent failed-proof hints | Shared Training DB / reference evidence → Rejection/Acceptance Logs → User Upload Files |
| Autonomous topic/title/metadata agents | User research prompt, compact metadata/candidate lists | Usually none; metadata is capped direct context | No RAG unless the specific browsing/expansion step calls for full-content fallback |
| Autonomous brainstorm/paper context | User research prompt, topic/title/outline as applicable | Brainstorm DB when the compiler caller injects it, current paper | Reference and prior-brainstorm papers are indexed as high-priority RAG evidence; oversized direct source blocks fail or use the mode's explicit fallback |
| Autonomous proof verification | User prompt, complete source content, proof candidate/formalization context | Verified proof summaries and failed-proof hints | Proof agents operate outside Chroma RAG; Lean source files are not indexed |
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.
- 9d ago First seen · 268 lines · 19 tokens per session scan A 71b6caa76833
rag-design-for-overall-program is a cursor rule published in the GitHub repository Intrafere/MOTO-Autonomous-ASI (83 stars, last pushed 5d ago), licensed MIT. It adds 19 tokens to every session and 5,746 once invoked, about $0.0001 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
cardano-mcp
These rules complement our main development collaboration rules, focusing specifically on the Model Context Protocol (MCP) server project with TypeScript. This project serves as a specialized RAG (Retrieval-Augmented Generation) gateway to existing Cardano resources, documentation, and tools to assist with dApp…
005_Dynamo_KV_Cache_Management_KVBM
NVIDIA Dynamo KV Cache Management (KVBM) - Advanced memory management and offloading strategies for optimal performance.
ponytail
Ponytail, lazy senior dev mode. Always pick the simplest solution that works.
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.
cli-error-handling
CLI command error handling patterns.