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/Veightor/chutes-agent-toolkitnpx agentmods add skills/veightor/chutes-agent-toolkit/chutes-teeWrote 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/veightor/chutes-agent-toolkit/chutes-tee)<a href="https://agentmods.dev/skills/veightor/chutes-agent-toolkit/chutes-tee"><img src="https://agentmods.dev/badge/skills/veightor/chutes-agent-toolkit/chutes-tee/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/veightor/chutes-agent-toolkit/chutes-tee"><img src="https://agentmods.dev/badge/skills/veightor/chutes-agent-toolkit/chutes-tee.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00079 | $0.00609 |
| Opus 5 | $0.00039 | $0.00304 |
| Sonnet 5 | $0.00016 | $0.00122 |
| Haiku 4.5 | $0.00008 | $0.00061 |
Grade A, and why
chutes-tee 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 9d 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.
What it actually says
Chutes TEE for Hermes
Status: SHAPE-VALID, verified live 2026-04-13 against a real TEE chute. The current scripts parse and validate evidence shape; full cryptographic validation requires Intel DCAP tooling. Full skill lives at
plugins/chutes-ai/skills/chutes-tee/.
When to use this skill
Use when a Hermes user wants to:
- Confirm whether a Chutes model/chute is confidential-compute backed.
- Fetch
/chutes/{id}/evidenceand inspect the evidence envelope. - Parse TDX v4 quotes or NVIDIA Hopper GPU attestation fields.
- Understand the limits of TEE: it helps isolate runtime execution, but does not automatically prove model weights, prompts, or downstream systems are safe.
Hermes-facing workflow
Run from the repo root:
python plugins/chutes-ai/skills/chutes-tee/scripts/fetch_evidence.py --chute-id <id> --out /tmp/evidence.json
python plugins/chutes-ai/skills/chutes-tee/scripts/verify_quote.py --evidence /tmp/evidence.json
python plugins/chutes-ai/skills/chutes-tee/scripts/verify_gpu_attestation.py --evidence /tmp/evidence.json
python plugins/chutes-ai/skills/chutes-tee/scripts/attest_chute.py --chute-id <id>
For model selection, still use the live models endpoint and filter confidential_compute: true.
Deep references
plugins/chutes-ai/skills/chutes-tee/SKILL.mdplugins/chutes-ai/skills/chutes-tee/references/evidence-envelope.mdplugins/chutes-ai/skills/chutes-tee/references/tdx-quote-anatomy.mdplugins/chutes-ai/skills/chutes-tee/references/dcap-trust-root.mdplugins/chutes-ai/skills/chutes-tee/references/what-tee-does-not-protect.md
Safety rules
- Do not claim cryptographic verification unless DCAP validation has actually run and passed.
- Do not rely solely on a
-TEEmodel-name suffix; use live metadata and evidence. - Keep evidence files out of commits unless intentionally adding sanitized fixtures.
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.
- 9d ago First seen · 51 lines · 79 tokens per session scan A dd5516a0ddb1
chutes-tee is a skill published in the GitHub repository Veightor/chutes-agent-toolkit (5 stars, last pushed today), licensed MIT. It adds 79 tokens to every session and 609 once invoked, about $0.0004 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 skills, from other repositories
chutes-media
This skill should be used when the user asks to "generate an image", "create a hero image", "make a video", "generate music", "create a voiceover", "text to speech", "edit this image", or otherwise produce image/video/music/speech assets inside a project. It documents the describe→generate workflow of the…
error-tracking-upload-source-maps-vite
Upload source maps to PostHog Error Tracking for Vite.
contributor-pipeline-gardening
Maintenance of the contributor issue pipeline for JSONbored/metagraphed — closing issues that are already done but not marked so, and keeping the contributor-available backlog at its 50-100+ steady-state floor with well-scoped new issues. Runs every 8h via the scheduled task (raised from daily on 2026-07-15 so the…
mcp-analytics
Add PostHog MCP analytics to a TypeScript, JavaScript, or Python MCP server. Instruments the server so every tool call, agent intent, and failure is captured as a $mcp event — the @posthog/mcp Node SDK for TS/JS, or posthog.mcp (shipped inside the posthog package) for Python. Detects the server style (official MCP SDK…
frontend-visual-qa
Systematic visual QA sweep of the live metagraphed frontend (apps/ui, served at metagraph.sh) — actually browsing pages at real viewports and looking for defects only visible on inspection (crowding, overflow, truncation, misalignment), not catchable by reading code or trusting a PR's own screenshot table. Produces…
bittensor
Use when a developer asks what a Bittensor subnet does, whether it's up right now, how to call/integrate its API, or what mining/validating one costs and earns — e.g. "which subnet does image generation", "is subnet 7 healthy", "call the Beam API for me", "does mining subnet 3 need a GPU", "what does the median miner…