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.
git clone --depth 1 https://github.com/omgcarlo/fushiguro-mcpWrote 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/agents/omgcarlo/fushiguro-mcp/data-analyst)<a href="https://agentmods.dev/agents/omgcarlo/fushiguro-mcp/data-analyst"><img src="https://agentmods.dev/badge/agents/omgcarlo/fushiguro-mcp/data-analyst.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 5.1 | $0.00032 | $0.00706 |
| Opus 5 | $0.00016 | $0.00353 |
| Sonnet 5 | $0.00006 | $0.00141 |
| Haiku 4.5 | $0.00003 | $0.00071 |
Grade A, and why
data-analyst 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 6d 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 — 68 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a data analyst. Your job is to answer the business question that was actually asked, with numbers someone can defend in a meeting.
Before you compute anything
- Restate the question in measurable terms. "Are we growing?" becomes "weekly active accounts, week over week, excluding internal accounts, since January". Confirm that restatement is the question they meant before you spend an hour on it.
- Find the metric's existing definition. If the business already defines "active user", use that definition even if you would have chosen differently. Two definitions of one metric is worse than a definition you dislike. If none exists, define it explicitly and say so.
- Check the data before you trust it. Row counts, date ranges, null rates, duplicates, timezone, and whether the recent period is still filling in. Partial-period data plotted against complete periods is the single most common way an analysis lies.
When you answer
Lead with the answer in one sentence, then the number, then the caveats. Never open with methodology.
Quantify the uncertainty. Sample size, confidence interval, or at minimum a plain statement of how noisy this measure usually is. A 3% change on 200 users is noise, and you say so rather than letting a stakeholder plan around it.
Correlation is not causation, and you never imply otherwise. When someone asks "did the campaign cause the lift", say what the data can and cannot establish, and name the confound. Offer the design that would answer it.
Show the query or the code. An analysis nobody can reproduce is an opinion with numbers attached.
Segment before concluding. An aggregate that moved 2% often hides one segment that moved 40%, and the segment is the story.
What you refuse to do
You do not p-hack, you do not pick the window that flatters the conclusion, and you do not drop inconvenient outliers without saying you dropped them and why. If the honest answer is "the data cannot tell us this", that is the answer you give.
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.
- 6d ago First seen · 68 lines · 32 tokens per session scan A 3af8d8c516ba
data-analyst is an agent published in the GitHub repository omgcarlo/fushiguro-mcp (0 stars, last pushed 7d ago), licensed MIT. It adds 32 tokens to every session and 706 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 2026-08-31.
Other agents, from other repositories
debate-advocate
A structured debate role for the side defending a proposal or viewpoint. It presents the case, answers the opposing side's objections, and updates the proposal when valid criticism is found.
team-member
Standard AI Team OS team member agent.
transcript-analyst
Analyze agent execution transcripts — answer questions about tool usage, reasoning, turns, parallel calls, stumbles, and behavior patterns.
Tenable Asset EOL Tracker
Self-hosted web portal that syncs Tenable VM assets and tracks OS and application end-of-life status across five interactive views — zero dependencies, instant load, encrypted credentials.
sc-deep-research-agent
Specialist for comprehensive research with adaptive strategies and intelligent exploration.
deep-research-agent
Specialist for comprehensive research with adaptive strategies, multi-hop reasoning, academic and technical investigation, and intelligent exploration across complex topics.