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 skills add ehlyzov/skills --skill skill-workbenchgit clone --depth 1 https://github.com/ehlyzov/skillsWrote 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/ehlyzov/skills/skill-workbench)<a href="https://agentmods.dev/skills/ehlyzov/skills/skill-workbench"><img src="https://agentmods.dev/badge/skills/ehlyzov/skills/skill-workbench/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/ehlyzov/skills/skill-workbench"><img src="https://agentmods.dev/badge/skills/ehlyzov/skills/skill-workbench.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 | $0.00058 | $0.01164 |
| Opus 5.5 | $0.00023 | $0.00466 |
| Sonnet 5.5 | $0.00012 | $0.00233 |
| Haiku 4.5 | $0.00006 | $0.00116 |
Grade A, and why
skill-workbench 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 15d 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.
This is a copy
100% identical to skill-workbench — 0 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 60 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Skill Workbench
Turn the user's workflow into a skill whose behavior can be tested and explained from artifacts. Write user-facing material in the user's language.
1. Establish the contract
Read the current request, prior decisions, and existing skill if supplied. Reuse answers already given. For an existing skill, inspect a reported failure before proposing a rewrite.
Capture a short brief outside the target skill:
- task, intended users, activation examples and exclusions;
- input and output contracts, missing-data behavior;
- observable success criteria and unacceptable outcomes;
- permitted actions and consequential decisions still unresolved.
Distinguish user requirements, source facts, and your proposals. Ask only questions whose answers materially change correctness, scope, or the consequences of an error. Batch a few focused questions with suggested defaults; do not run an exhaustive interview. Decide routine wording, file layout, and implementation yourself. Explicit user delegation allows you to choose policy; record it as delegated, not as a supplied requirement.
Delegation of routine implementation is not delegation of interpretation rules. Rules that change which records qualify, who owns an action, or how an ambiguous deadline becomes a date affect the result. If the user has not specified or explicitly delegated them, ask a focused question before finalizing that behavior. Do not call a plausible convention a technical detail.
If a consequential criterion is unknown, ask and continue independent preparation. Do not silently convert an assumption into a business rule. A blocked draft is useful; a falsely finished skill is not.
2. Define examples before implementation
Use the eval procedure. An eval is an input plus observable expectations fixed before execution, not a second prompt asking whether the first answer is good.
For a new skill, start with a normal case, a missing/invalid-input case, and a boundary case when relevant. Include activation and non-activation examples. User-provided examples are preferable; mark invented examples as synthetic and expectations as author-generated. Reserve fresh cases for subsequent evaluation. For a narrow fix, the reproducer and relevant regression cases may suffice.
What ships with it
4 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 15d ago First seen · 60 lines · 58 tokens per session scan A bb98cbc1854a
skill-workbench is a skill published in the GitHub repository ehlyzov/skills (2 stars, last pushed 16d ago), licensed MIT. It adds 58 tokens to every session and 1,164 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 0 findings. It is 100% identical to skill-workbench, differing in 0 lines, and is treated as a copy.
Other skills, from other repositories
research-engineer
An uncompromising Academic Research Engineer. Operates with absolute scientific rigor, objective criticism, and zero flair. Focuses on theoretical correctness, formal verification, and optimal implementation across any required technology.
tika-eval-compare
Compare extracts from two Tika builds over a corpus to detect regressions in content, encoding, exceptions, and embedded-document handling. Use for "compare before/after extracts", "eval this change against the corpus".
cua-driver
Use Cua Driver for desktop or browser tasks that are awkward or unavailable through Bash/APIs, or when the user explicitly wants GUI interaction: app testing, visual bug reproduction, form filling, calendar entry, screenshots, and demo recording. Also covers Cua setup; not OpenAI Codex Computer Use or web research.
jetson-validate-image
Use after jetson-flash-image to run static BSP checks, on-target smoke/regression tests on a flashed DUT, or both. Not for build or flash steps. Triggers: validate bsp, on-target validation.
atmos-validation
Validate Atmos projects, components, arbitrary JSON Schema inputs, EditorConfig, and GitHub Actions; use affected-file selection and native CI annotations.
skill-benchmark
Benchmark AI skill effectiveness by measuring implementation quality against legacy constraints.