ComfyUI-AIAssistant AGENTS.md

ComfyUI-AIAssistant AGENTS.md is an instructions file for Codex, OpenCode from alksef/ComfyUI-AIAssistant. It costs 399 tokens per session, scanned A, original, MIT.

Repository instructions for ComfyUI-AIAssistant, a ComfyUI extension that gives a local AI assistant a limited, read-only view of the selected canvas area. They also define one guarded exception for writing text into a selected text-like widget.

In plain words
What is it for?
Use them when developing or reviewing this ComfyUI extension, especially its context API, MCP endpoint, guarded widget editing, tests, and verification workflow.
Why use it?
They clarify what the assistant may access or change, helping developers avoid modifying workflows, files, settings, or other state by mistake. They also set rules for privacy, bounded data, logging, and dependencies.

Instructions file for CodexOpenCode

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 instructions/alksef/comfyui-aiassistant/agents-md
Clone the repo
git clone --depth 1 https://github.com/alksef/ComfyUI-AIAssistant

Made for: Codex, OpenCode.

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 ComfyUI-AIAssistant AGENTS.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/alksef/comfyui-aiassistant/agents-md.svg)](https://agentmods.dev/instructions/alksef/comfyui-aiassistant/agents-md)
Your own site
<a href="https://agentmods.dev/instructions/alksef/comfyui-aiassistant/agents-md"><img src="https://agentmods.dev/badge/instructions/alksef/comfyui-aiassistant/agents-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 399 This file is loaded in full into every session.
When invoked 399 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 $0.00399 $0.00399
Opus 5 $0.00199 $0.00199
Sonnet 5 $0.00080 $0.00080
Haiku 4.5 $0.00040 $0.00040

Measured 3d ago against content hash 4390fda0f412, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

ComfyUI-AIAssistant AGENTS.md 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 3d 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.

AGENTS.md · 48 lines

What it actually says

ComfyUI-AIAssistant agent guide

Scope

This repository contains a small ComfyUI extension that exposes a bounded, read-only snapshot of the user's current canvas selection to a local AI assistant.

Boundaries

  • Do not add executable ComfyUI workflow nodes.
  • Do not mutate the graph, widgets, queue, models, files, or ComfyUI settings. Single explicit carve-out (ROADMAP-003): at an agent's MCP request the plugin may write text into exactly one text-like widget of the currently selected node on the active page, guarded by revision and page-identity CAS and size bounds; the edit behaves like a manual widget edit.
  • Do not persist workflow, prompt, selection, or conversation content.
  • Do not make outbound network requests. Consumers pull context from the ComfyUI localhost API.
  • The MCP endpoint POST /mcp is one more inbound route beside the context routes: beyond the guarded write tool it never mutates state, never logs payloads, and adds no dependencies.
  • Keep runtime dependencies empty. Use APIs already shipped with ComfyUI.
  • Keep the public context contract versioned and bounded.
  • Never log the context payload or prompt/widget contents.

Development workflow

  • Durable decisions belong in docs/roadmaps/.
  • Non-trivial implementation starts from a written task: goal, allowed files, verified facts, required behavior, acceptance commands.
  • Review every diff before accepting it and run the task acceptance commands.
  • Do not commit unless the owner explicitly asks.

Verification

Run from the repository root:

python -m unittest discover -s tests -v   # backend (needs aiohttp)
ruff check ai_assistant tests
ruff format --check ai_assistant tests
npm test                                  # frontend (node:test)
node --check web/aiAssistant.js
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. 3d ago First seen · 48 lines · 399 tokens per session scan A 4390fda0f412

Subscribe to this mod's changes

ComfyUI-AIAssistant AGENTS.md is an instructions file published in the GitHub repository alksef/ComfyUI-AIAssistant (0 stars, last pushed 6d ago), licensed MIT. It adds 399 tokens to every session, about $0.0020 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.