Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/artokun/comfyui-mcpnpx agentmods add skills/artokun/comfyui-mcp/triton-sageattentionWrote 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/skills/artokun/comfyui-mcp/triton-sageattention)<a href="https://agentmods.dev/skills/artokun/comfyui-mcp/triton-sageattention"><img src="https://agentmods.dev/badge/skills/artokun/comfyui-mcp/triton-sageattention/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/skills/artokun/comfyui-mcp/triton-sageattention"><img src="https://agentmods.dev/badge/skills/artokun/comfyui-mcp/triton-sageattention.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 2 findings, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Tool Misuse · line 328 Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).Fix: Validate all tool parameters against an allowlist. Reject dangerous parameter values (shell=True, --force, -rf /) and use safe defaults.
- high Tool Misuse · line 328 Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).Fix: Validate all tool parameters against an allowlist. Reject dangerous parameter values (shell=True, --force, -rf /) and use safe defaults.
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 | $0.00166 | $0.05382 |
| Opus 5 | $0.00083 | $0.02691 |
| Sonnet 5 | $0.00033 | $0.01076 |
| Haiku 4.5 | $0.00017 | $0.00538 |
Grade C, and why
triton-sageattention scanned grade C with 1 finding 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 13d 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.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
rm -rf ~/.triton Copies of this mod
1 near-identical copy found in the catalogue:
- triton-sageattention — 86% identical, 270 lines differ
How it starts
The opening of the file, as written. The whole thing — 380 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Triton + SageAttention (ComfyUI acceleration)
See also
comfyui-launch-flagsfor the full attention / VRAM / cache flag matrix. Note the Z-Image exception: Z-Image is broken under--use-sage-attention, so launch it with--use-pytorch-cross-attentioninstead.
Prefer kitchen INT8 attention when it is available
If kitchen action:"status" (or panel_kitchen) reports kitchen present and
int8_attention_is_available on this GPU, launch with --use-ck-attention
and skip the sageattention wheel dance. Kitchen INT8 attention is a ComfyUI
flag; it does not need a version-matched sageattention wheel. Restart
required, consent-gated like every restart.
Only fall through to the Triton + SageAttention install below when kitchen INT8 is unknown or not available. A failed kitchen probe is unknown, not a no.
Overview
Two optional accelerators that many modern video graphs (especially kijai's ComfyUI-WanVideoWrapper) reference by default:
- SageAttention (
import sageattention), a quantized attention kernel. Selected via a node'sattention_mode = sageattn(WanVideoWrapper) or ComfyUI's--use-sage-attentionstartup flag. ~20 to 40% faster sampling on supported NVIDIA GPUs. - Triton, the GPU kernel compiler that inductor
torch.compileneeds. WanVideoWrapper'sWanVideoTorchCompileSettings(and anytorch.compile/ inductor node) compiles the model through Triton for another speedup.
The risk. Both are version-locked to your exact torch + CUDA + python. A wrong wheel does worse than fail to install. It can break the torch install (mismatched CUDA DLLs,
ImportError, or silent NaNs). And the failure mode of not having them is a hard crash before any sampling:ValueError: Can't import SageAttention: No module named 'sageattention', or compile errors /triton: unavailablein the startup log. This is exactly thevideo-extendTRAP 5.
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.
- 13d ago First seen · 380 lines · 166 tokens per session scan C 6b6502b2104a
triton-sageattention is a skill published in the GitHub repository artokun/comfyui-mcp (740 stars, last pushed 2d ago), licensed MIT. It adds 166 tokens to every session and 5,382 once invoked, about $0.0008 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
chief-revenue-officer
Owns the revenue engine end to end: sales, monetization, pricing, customer success, retention, and partnerships. Use this for pricing and packaging decisions, sales strategy and coverage, forecast and pipeline health, churn and expansion, partner and channel strategy, or when marketing-sourced demand is not…
chief-strategy-officer
Owns where the business plays and how it wins over a multi-year horizon — portfolio choices, corporate development, strategic partnerships, and planning under uncertainty. Use this for a decision about which markets or businesses to be in, whether to build, buy, or partner, how to allocate capital across business…
dev
One-command dev loop boot. Spins up portless (named HTTPS subdomain), emulate (stateful API mocks), the project's dev server, and an agent-browser session, all keyed to the current git branch. Use when starting a feature branch, switching worktrees, or returning to a project after a break. Skips silently with install…
storybook-mcp-integration
Reference for the Storybook MCP server itself (@storybook/addon-mcp): 6 tools across 3 toolsets (dev, docs, testing), availability detection, and per-agent toolset filtering. Use when setting up the server or calling these tools directly against components that already exist. For the end-to-end pipeline that turns a…
Product Photoshoot Workflow
A workflow for turning one product photo into a set of four e-commerce images from different visual directions.
neo4j-modeling-skill
Design, review, and refactor Neo4j graph data models. Use when choosing node labels vs relationship types vs properties, migrating relational/document schemas to graph, detecting anti-patterns (generic labels, supernodes, missing constraints), designing intermediate nodes for n-ary relationships, enforcing schema with…