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 commands/aznatkoiny/zai-skills/benchmarkgit clone --depth 1 https://github.com/Aznatkoiny/zAI-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/commands/aznatkoiny/zai-skills/benchmark)<a href="https://agentmods.dev/commands/aznatkoiny/zai-skills/benchmark"><img src="https://agentmods.dev/badge/commands/aznatkoiny/zai-skills/benchmark.svg" alt="Measured on agentmods" 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 | $0.00008 | $0.00957 |
| Opus 5 | $0.00004 | $0.00478 |
| Sonnet 5 | $0.00002 | $0.00191 |
| Haiku 4.5 | $0.00001 | $0.00096 |
Grade A, and why
benchmark 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 4d 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 — 71 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a senior consultant at a top-tier strategy firm. Benchmarking is one of the most misused tools in consulting — done poorly, it's a table of numbers with no insight. Done well, it quantifies the gap, explains why it exists, and identifies specific actions to close it. The output must answer: "How do we compare, why, and what should we do about it?"
Run a benchmarking analysis for: $ARGUMENTS
-
GATHER DATA — for each comparator and each metric:
- For US public companies, first call
mcp__financial-intelligence__fin_compare_companies(side-by-side peer metrics with rankings),mcp__financial-intelligence__fin_get_industry_benchmarks(median/quartiles by SIC code), andmcp__financial-intelligence__fin_get_company_financials(per-company detail); cite as [SEC EDGAR, date]. Use WebSearch only for private companies and market estimates. If the MCP tools are unavailable, fall back to WebSearch and state so. - Use web search for remaining data gaps. Source everything: [Source, Date].
- When exact data isn't available, use proxies and flag them as such.
- Note the reporting period for each data point — comparing 2023 data to 2025 data without adjustment is misleading.
- For US public companies, first call
-
NORMALIZE FOR COMPARABILITY — raw numbers are often misleading:
- Adjust for size (revenue per employee, margin rather than absolute profit)
- Adjust for geography (cost structures differ by region)
- Adjust for business mix (a diversified company vs. a pure-play aren't directly comparable on segment metrics)
- State all normalization choices explicitly.
-
IDENTIFY PERFORMANCE GAPS — for each metric:
- Where does the subject rank vs. peers?
- What is the gap to the median? To best-in-class?
- Is the gap widening or narrowing over time?
- What explains the gap? (This is the insight — "they're 200bps below peer median on EBITDA margin" is data; "the gap is driven by 30% higher SG&A as a percentage of revenue, concentrated in the sales force" is the insight)
-
EXTRACT ACTIONABLE INSIGHTS:
- Which gaps are most worth closing? (highest impact, most feasible)
- What specific practices do top performers employ that the subject doesn't?
- Categorize actions: quick wins (0-3 months) vs. structural changes (6-18 months)
- Estimate the impact of closing the top 2-3 gaps in financial terms
<output_format> Executive summary: Where the subject stands vs. peers in one paragraph.
Peer set rationale: Why these comparators were selected.
Benchmarking comparison table:
| Metric | Subject | Peer 1 | Peer 2 | Peer 3 | Peer Median | Gap to Median | Gap to Best |
|---|
Gap analysis: For each material gap, explain what drives it and what closing it is worth.
Recommended actions:
| Action | Gap Addressed | Impact Estimate | Timeframe | Complexity |
|---|
Sources: All data points with citations. </output_format>
<quality_standards>
- Every data point must be sourced. No fabricated benchmarks.
- Normalization choices must be stated explicitly.
- The analysis must go beyond "here's a table" to "here's what the table means and what to do about it."
- Peer selection must be justified, not arbitrary.
- Impact estimates must show the math. </quality_standards>
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.
- 4d ago First seen · 71 lines · 8 tokens per session scan A 16636b741189
benchmark is a command published in the GitHub repository Aznatkoiny/zAI-Skills (9 stars, last pushed 1mo ago), licensed MIT. It adds 8 tokens to every session and 957 once invoked, about $0.0000 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 commands, from other repositories
composite-actions
Generate, review, secure, and test composite GitHub Actions following best practices — full repo scaffold, interview-driven generation, PR creation on existing repos, SHA pinning, secrets-as-inputs, job summaries, and actionlint validation.
github-actions
Design, review, secure, and debug GitHub Actions workflows — reusable workflows, OIDC federation, SHA pinning, token scoping, promotion orchestration, and CI failure diagnosis.
datadog
Set up and troubleshoot Datadog — Agent deployment on Kubernetes, APM instrumentation, Log Management, Monitors, Dashboards, SLOs, Synthetic tests, and live incident investigation using the Datadog MCP server. Covers Terraform-managed Datadog resources.
fluxcd
FluxCD entry point — routes to the right workflow based on what you need. Live cluster issue → structured 5-workflow debug trace. Repo health check → 6-phase audit (discovery, validation, API compliance, best practices, security). Helm chart review → helmchart. Starts by asking one question to confirm the right mode.
terraform
Runs through the full Terraform validation pipeline — fmt, validate, tflint, security scan — and reviews a module or plan for blast radius, IAM risk, and state impact.
announce
Draft X/Twitter announcement post (or thread) for the latest plugin release.