skill-benchmark-design

skill-benchmark-design is a skill for Codex from jeremylongworth-source/AgentSkills. It costs 47 tokens per session (520 once invoked), scanned A, original, MIT.

A plan for building small, realistic benchmarks that test whether an agent skill improves its results.

In plain words
What is it for?
Designing normal, edge-case, risky, and repeated-use scenarios, capturing outputs and validation results, and deciding whether to keep, revise, split, merge, defer, or retire a skill.
Why use it?
It makes skill evaluation repeatable by defining the same inputs, scoring rules, evidence, and decision criteria for baseline and skill-assisted runs.

Skill for Codex

Written for Codex: agents/openai.yaml present. Also seen: mentions Claude Code; mentions Codex.

Good fit Designing normal, edge-case, risky, and repeated-use scenarios, capturing outputs and validation results, and deciding whether to keep, revise, split, merge, defer, or retire a skill.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/jeremylongworth-source/agentskills/skill-benchmark-design
Install

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.

Any agent
npx skills add jeremylongworth-source/AgentSkills --skill skill-benchmark-design
Clone the repo
git clone --depth 1 https://github.com/jeremylongworth-source/AgentSkills

Made for: Codex.

Wrote 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.

agentmods badge for skill-benchmark-design

README.md
[![agentmods](https://agentmods.dev/badge/skills/jeremylongworth-source/agentskills/skill-benchmark-design/github.svg)](https://agentmods.dev/skills/jeremylongworth-source/agentskills/skill-benchmark-design)
Your own site
<a href="https://agentmods.dev/skills/jeremylongworth-source/agentskills/skill-benchmark-design"><img src="https://agentmods.dev/badge/skills/jeremylongworth-source/agentskills/skill-benchmark-design/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.

agentmods 80×15 button for skill-benchmark-design

Your own site · 80×15
<a href="https://agentmods.dev/skills/jeremylongworth-source/agentskills/skill-benchmark-design"><img src="https://agentmods.dev/badge/skills/jeremylongworth-source/agentskills/skill-benchmark-design.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 47 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 520 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00047 $0.00520
Opus 5 $0.00023 $0.00260
Sonnet 5 $0.00009 $0.00104
Haiku 4.5 $0.00005 $0.00052

Measured 8d ago against content hash 5a4099bb5627, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

skill-benchmark-design 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 8d 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.

skills/skill-benchmark-design/SKILL.md · 65 lines

How it starts

The opening of the file, as written. The whole thing — 65 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Skill Benchmark Design

Core Workflow

  1. Name the skill or bundle under test, target user, target artifact, and decision the benchmark must support.
  2. Select realistic scenarios that exercise normal, edge, and risky requests without leaking expected answers into the prompt.
  3. Define baseline and skill-enabled runs using the same prompt, source material, and acceptance criteria.
  4. Choose scoring criteria from docs/evaluation/skill-quality-rubric.md and add task-specific pass/fail checks.
  5. Specify evidence capture: prompt, inputs, outputs, reviewer notes, validation output, loaded context, tools/scripts, and overhead when available.
  6. Set the decision rule: keep, revise, split, merge, defer, or retire.

Overhead And Composition Checks

  • Include one repeated-run or batch scenario when the skill may be used many times in one thread.
  • Compare monolithic guidance against composed skills or skillsets when the workflow has separable stages.
  • Capture activated skills, loaded references, command/script output, MCP tool use, artifacts, and validation results.
  • Keep host optimizations optional. Benchmark Claude Code context: fork, dynamic !command injection, or MCP Tool Search separately from the portable AgentSkills baseline.
  • Do not report exact context or token savings unless the host exposes reliable measurements.

Safety Rules

  • Do not use private customer, employee, security, or financial data in public benchmark artifacts.
  • Do not optimize prompts by exposing the expected answer.
  • Do not treat an automated judge as sufficient for legal, financial, security, employment, medical, or production-impacting workflows.
  • Keep tracing, scoring, and observability tools optional and vendor-neutral.

Deliverable Shape

For benchmark plans, provide:

  • Skill or bundle under test
  • Target artifact and target user
  • Scenarios and source material
  • Baseline and skill-enabled run plan
  • Acceptance criteria and scoring rubric
  • Evidence capture plan
  • Overhead and composition checks
  • Safety boundaries
  • Promotion decision rule

Read the full file on GitHub · 65 lines

Files

What ships with it

2 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.

Changes

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.

  1. 8d ago First seen · 65 lines · 47 tokens per session scan A 5a4099bb5627

Subscribe to this mod's changes

skill-benchmark-design is a skill published in the GitHub repository jeremylongworth-source/AgentSkills (1 stars, last pushed 9d ago), licensed MIT. It adds 47 tokens to every session and 520 once invoked, about $0.0002 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 2026-09-03.