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/beingmartinbmc/jambavanWrote 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/beingmartinbmc/jambavan/jambavan-mcp)<a href="https://agentmods.dev/rules/beingmartinbmc/jambavan/jambavan-mcp"><img src="https://agentmods.dev/badge/rules/beingmartinbmc/jambavan/jambavan-mcp.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 5.1 | $0.01026 | $0.01026 |
| Opus 5 | $0.00513 | $0.00513 |
| Sonnet 5 | $0.00205 | $0.00205 |
| Haiku 4.5 | $0.00103 | $0.00103 |
Grade A, and why
jambavan-mcp 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 6d 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 — 50 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Jambavan MCP
Jambavan is a local-first code index, project memory, failure ledger, and review assistant. It does not call an LLM and never uploads code. Use its MCP tools when they are available, and never invent a Jambavan result when the server is not.
Requires jambavan >= 2.0.1. If tools/list still advertises jambavan_graph_report or jambavan_memory_recall, the host is pinned to a 1.x build — say so instead of guessing which names work.
Session start
jambavan_doctor {}first — readProject rootand itssourcebefore anything else:- Root is the repo you are working in → go to step 2.
source: cwd-fallback— the root fell back to a home directory rather than a project, so every repository-bound tool is blocked (jambavan_context,jambavan_index,jambavan_graph,jambavan_impact,jambavan_review_pack, and the file tools). Bind withjambavan_awaken { "root": "<absolute path to this repo>" }.- Root is a different repo → whether you can claim it depends on that root's
source; see the next section.
jambavan_index {}when the index is missing or stale, thenjambavan_watch { "action": "start" }if the session will continue past a single edit.jambavan_awaken {}for the operating protocol and memory scope. Its output supersedes this rule if the two disagree.
One server, many workspaces
Some hosts run a single Jambavan process for every open window — Cursor is one — so the project root is shared rather than per-workspace.
A root that a model set through jambavan_awaken or jambavan_index (source: tool-input) can be re-pointed: bind the repo you want and the server stops the old watcher, closes the old index, and follows you. Two projects taking turns is normal and cheap once each has an index.
A root that came from JAMBAVAN_ROOT (source: env), the host's roots/list (client-roots), or the server's own working directory (cwd-project) is immutable, and binding a different one fails with root is already fixed by .... Do not retry that call. Either report that Jambavan is unavailable for this repo and fall back to host tools, or give the repo its own MCP entry with JAMBAVAN_ROOT.
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.
- 6d ago First seen · 50 lines · 1,026 tokens per session scan A 9c987a1c3e67
jambavan-mcp is a cursor rule published in the GitHub repository beingmartinbmc/jambavan (2 stars, last pushed 3d ago), licensed MIT. It adds 1,026 tokens to every session, about $0.0051 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 2026-08-31.
Other cursor rules, from other repositories
openviking-memory
Use OpenViking as the durable memory and context backend.
vectr
Vectr tool usage rules for AI-assisted development.
memento-conventions
Memento OS — reasoning artifact conventions for persistent AI memory.
avos-agent-workflow
Avos memory workflow for AI coding agents. Guides codebase search and history lookup.
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.