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.
npx agentmods add skills/rust-works/succinctly/benchmark-docsnpx skills add rust-works/succinctly --skill benchmark-docsgit clone --depth 1 https://github.com/rust-works/succinctlyWhat 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 | $0.00051 | $0.03047 |
| Opus 5 | $0.00026 | $0.01523 |
| Sonnet 5 | $0.00010 | $0.00609 |
| Haiku 4.5 | $0.00005 | $0.00305 |
Grade A, and why
benchmark-docs 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 2d 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.
How it starts
The opening of the file, as written. The whole thing — 299 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Benchmark Documentation Skill
This skill ensures proper handling of benchmark documentation across multiple platforms (ARM/Apple Silicon and x86_64/Intel/AMD).
For comprehensive benchmarking instructions, see docs/guides/benchmarking.md.
This skill focuses on documentation-specific rules and multi-platform considerations.
A/B Measurement Rules (before/after a code change)
Full method and evidence: docs/guides/benchmarking.md § A/B Benchmarking Method. These seven rules each cost a wrong conclusion on #106 — the naive method reported a 16x win as a regression.
- Interleave the two binaries within each repetition. Never run all of A then all of B — the second half starts thermally loaded, which made an improved binary measure up to 2x slower on every workload. This is a design fix; more reps do not help, and min/median can agree with each other while both are wrong.
- Process-spawn A/B needs inputs >= 1 MB. Startup is ~4-6 ms, i.e. the whole runtime at
1kb/10kb/100kb — which are in
dev bench yq's defaults. Use--sizes 1mb,10mb. - Report the scaling curve across 2-3 sizes. A speedup that grows with input size is the signature of an algorithmic fix and corroborates the claimed mechanism; one ratio does not.
- Gate on output identity first. Diff both binaries over every input x query and confirm
they still match
jq/yq. A faster binary that changed behaviour is not a win. - Measure both architectures — the effect size differs, not just the noise. #106 was 6.1x on M4 Pro and 16.4x on Zen 4 for the same commit, because cache-bound costs do not port.
- Verify the box is idle, and distrust macOS load average — it counts uninterruptible-wait
threads, reading 1.4 on a machine using 2.4% CPU. Sum
ps -Ao pcpu, check forcargo|rustc|claude, and require AC power. - A benchmark cannot measure a shape it does not generate. "All neutral" is not evidence if the suite lacks the relevant input; add the generator pattern first, then measure.
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.
- 2d ago First seen · 299 lines · 51 tokens per session scan A 63aff19e3262
benchmark-docs is a skill published in the GitHub repository rust-works/succinctly (51 stars, last pushed 2d ago), licensed MIT. It adds 51 tokens to every session and 3,047 once invoked, about $0.0003 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.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
babysit-pr
Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…
imagegen
Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
next-cache-components-optimizer
Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…