aide-device-benchmark-runner SKILL.md skillA
aide-device-benchmark-runner from anonymousnomad/aide-sovereign-workbench is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.
mdr add anonymousnomad/aide-sovereign-workbench/aide-device-benchmark-runner@git:20260823.df7cbdePins git:20260823.df7cbde in mdr.lock. Pin the fingerprint instead.
- Audit grade
- A, 17 of 17 checks passed
- Kind
- SKILL.md skill
- Repository
- anonymousnomad/aide-sovereign-workbench
- Path
skills/packs/aide-device-benchmark-runner/SKILL.md- Versions recorded
- 1
- Latest version
- git:20260823.df7cbde, committed Aug 23, 2026
- Size
- 7,146 bytes
- Repository stars
- 53
- License
- Apache-2.0
SOP for AIDE's MEASURED device benchmark lane — running llama.cpp llama-bench against a candidate GGUF on the user's actual hardware, parsing pp/tg tok/s from JSON, caching verdicts, and merging measured numbers with M2 header-estimate fit verdicts. Extends (does not replace) aide-model-hub-acquisition M2 estimates. Use when implementing the benchmark runner, when "Will this run fast?" must be answered with numbers instead of estimates, when interpreting llama-bench output, or when debugging…
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# Device Benchmark Runner — measured tok/s, not vibes
## What we are doing
Give AIDE a real benchmark capability: user picks a model (or finishes a download) -> daemon spawns llama.cpp's official `llama-bench` binary -> parses JSON output -> caches {pp512_tok_s, tg128_tok_s, stdev, backend, ngl, gpu_name, timestamp} per model+device -> UI shows "runs at ~X tok/s on YOUR machine" and feeds the recommender ranking.
## Research base (primary sources, verified 2026-08-24)
1. llama.cpp `tools/llama-bench/README.md` (ggerganov/llama.cpp master): three test types — pp (prompt processing), tg (text generation), pg (combined); defaults `-p 512 -n 128`, reps default 5; results = mean ± stdev tok/s; **measurements EXCLUDE tokenization and sampling**; `-o json|csv|md` output; `-ngl` controls GPU layer offload; `-fa` flash attention; multiple values via commas run combinatorial matrix.
…Read the whole file at its exact version.
How to install
mdr add anonymousnomad/aide-sovereign-workbench/aide-device-benchmark-runner@sha256:72096ccc7a328fbfThe line at the top of this page pins the label git:20260823.df7cbde, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
[](https://markdownregistry.com/a/art_34ye3gkj44jue63d)
4 badge image fetches in 30 days. What the badge shows.
Versions
The registry has recorded 1 version of this file, committed on Aug 23, 2026.
Audit of the latest version
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (7146 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
anonymousnomad/aide-sovereign-workbench · 53 stars · license Apache-2.0 · pushed 2026-10-26 · branch covert-production
API
GET https://markdownregistry.com/api/v1/artifacts/art_34ye3gkj44jue63d GET https://markdownregistry.com/api/v1/resolve?ref=anonymousnomad/aide-sovereign-workbench/aide-device-benchmark-runner GET https://markdownregistry.com/api/v1/blob/72096ccc7a328fbf20e27e25ff48b9e36eed4b3c2ea4a2b681f8e8b90ce4b813
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.