cursor

A Cursor rule that routes a project to generated instructions stored in protocol files. Cursor is a code editor that can apply project-specific rules while an agent works.

In plain words
What is it for?
Loading or refreshing project protocols, checking project status when requested, and retrieving additional instructions only when the work needs them.
Why use it?
It tells the agent where the current project’s instructions come from and what to do when those generated instructions are missing or outdated.

Cursor rule

Install

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.

agentmods
npx agentmods add rules/srmackey/insitu/cursor
Clone the repo
git clone --depth 1 https://github.com/srmackey/insitu
Per session 598 This file is loaded in full into every session.
When invoked 598 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00598 $0.00598
Opus 5 $0.00299 $0.00299
Sonnet 5 $0.00120 $0.00120
Haiku 4.5 $0.00060 $0.00060

Measured today against content hash f7d2c0eccada, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

cursor 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.

install/routers/cursor.mdc · 43 lines

What it actually says

Insitu router

Insitu stores reusable stanzas and composes a project protocol. The project key is the working folder basename (projects/<folder>/).

The composed core lives in this checkout as PROTOCOL.md and as the host protocol file under .cursor/rules/insitu-protocol.mdc. Treat that generated pack as binding. Do not edit it.

If the pack is missing, or its header is stale relative to the vault, call materialize (or resolve_protocol then materialize) before other work.

materialize writes only that generated pack. It never writes AGENTS.md or CLAUDE.md. If other host files this host loads are missing, retrieve the multi-platform pack (get_stanza platform/router-vs-pack and platform/cursor) and write them here. Do not put pack bodies in this router. If those stanzas miss, the pack is not installed; say so. Do not invent files the pack does not name.

If the user asks Insitu status of this folder, call project_status. Do not walk get_project, resolve_protocol, or get_role for that readout. Do not call it at session start.

Pull on-demand stanzas with get_stanza when the work needs one. Do not pull the on-demand list at session start. Use list_stanzas to see the catalog and sizes (what to include, what to trim). Do not call it at session start to load the core.

Mapped skills are generated copies under the host skill directories (.grok/skills, .claude/skills, .cursor/skills). Treat them as binding for /name. Do not edit them. If a generated skill is missing or its stamp is stale relative to the vault, call materialize. Pull a skill body for inspect with get_skill. Use list_skills to see the catalog and sizes (what to attach). Do not call list_skills at session start to load skills. Host discovery is the load path.

After an authoring write, if this working folder's project key is in affects_projects:

  1. Call materialize on this checkout.
  2. Tell the user the loaded protocol is stale and they should start a new session.
  3. Do not continue as if the new stanza or role is already in core.

If this project is not in the list, keep working.

Do not call delete_stanza, delete_role, delete_project, or delete_skill unless the user explicitly asked to delete that object. Do not treat validate findings as a reason to delete. Never clean up the vault.

This router is not the protocol. It does not contain project-specific guidance.

Changes

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.

  1. today First seen · 43 lines · 598 tokens per session scan A f7d2c0eccada

Subscribe to this mod's changes

cursor is a cursor rule published in the GitHub repository srmackey/insitu (0 stars, last pushed yesterday), licensed MIT. It adds 598 tokens to every session, about $0.0030 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.