agent-host-isolation

agent-host-isolation is a cursor rule for Cursor from csmar432/finai-research. It costs 21 tokens per session (322 once invoked), scanned A, original, MIT.

A set of operating rules for an AI coding agent when it must work without asking for confirmation. It also defines how to handle financial research, evidence gaps, mock data, and final deliverables.

In plain words
What is it for?
Use it for automated research runs, empirical analysis, data-source checks, and final handoffs that must include both completed work and documented limitations.
Why use it?
It prevents the agent from quietly substituting different data, inventing results, or presenting incomplete research as finished. It also keeps unattended work from stopping for user approval.

Cursor rule for Cursor

Written for Cursor: installed under .cursor/.

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/csmar432/finai-research/agent-host-isolation
Clone the repo
git clone --depth 1 https://github.com/csmar432/finai-research

Made for: Cursor.

Wrote 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.

agentmods badge for agent-host-isolation

README.md
[![agentmods](https://agentmods.dev/badge/rules/csmar432/finai-research/agent-host-isolation.svg)](https://agentmods.dev/rules/csmar432/finai-research/agent-host-isolation)
Your own site
<a href="https://agentmods.dev/rules/csmar432/finai-research/agent-host-isolation"><img src="https://agentmods.dev/badge/rules/csmar432/finai-research/agent-host-isolation.svg" alt="Measured on agentmods" height="20"></a>
Per session 21 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 322 The whole file, excluding the scripts and references it only reads on demand.
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 5.1 $0.00021 $0.00322
Opus 5 $0.00010 $0.00161
Sonnet 5 $0.00004 $0.00064
Haiku 4.5 $0.00002 $0.00032

Measured 6d ago against content hash 4adc37241840, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

agent-host-isolation 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.

.cursor/rules/agent-host-isolation.mdc · 24 lines

What it actually says

Agent-host isolation (when user says 不需要确认 / FINAI_NO_HITL / 交 FINAL)

When the user (or slot) requires no questions, no wait, no Mock:

  1. Prefer python scripts/agent_host_entry.py for writing; for empirics prefer python -m scripts.research_framework.enhanced_pipeline --explore --panel ... (or modern_did / robustness_runner). Thin official results are a product bug — deepen inside FinAI, don't invent run_real_*.py outside the APIs.
  2. Set / respect FINAI_EMPIRICAL_DATA_ROOT. Local panels before remote / proxy.
  3. Never substitute: city patents → firm green patents; overseas revenue → customs HS; interest coverage → bond spreads. Record hard-gaps in SKIPPED_CONFIG.md.
  4. Delivery must include FINAL.md + SKIPPED_CONFIG.md (not only CODEX_FINAL.md / PDF).
  5. Writing success with empirics gaps = partial, not journal-ready causal completion.
  6. Do not start MCP servers unless the user asked; do not invent coefficients.
  7. Demo/Mock panels require explicit --allow-demo / allow_synthetic=True.

Interactive research chats (no autopilot) still follow system-init.mdc greet + HITL.

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. 6d ago First seen · 24 lines · 21 tokens per session scan A 4adc37241840

Subscribe to this mod's changes

agent-host-isolation is a cursor rule published in the GitHub repository csmar432/finai-research (114 stars, last pushed yesterday), licensed MIT. It adds 21 tokens to every session and 322 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.