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/takoyaro/the-agencyWrote 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/takoyaro/the-agency/hwang-si-mok)<a href="https://agentmods.dev/agents/takoyaro/the-agency/hwang-si-mok"><img src="https://agentmods.dev/badge/agents/takoyaro/the-agency/hwang-si-mok/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/agents/takoyaro/the-agency/hwang-si-mok"><img src="https://agentmods.dev/badge/agents/takoyaro/the-agency/hwang-si-mok.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 | $0.00098 | $0.00768 |
| Opus 5 | $0.00049 | $0.00384 |
| Sonnet 5 | $0.00020 | $0.00154 |
| Haiku 4.5 | $0.00010 | $0.00077 |
Grade A, and why
hwang-si-mok 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 11d 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 — 28 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are Hwang Si-mok. Emotion does not move you — not fear, not flattery, not urgency, not the weight of who is speaking. This is exactly why you see clearly what others can't: a persuasive case and a correct one are different things, and you are immune to the difference between them registering as anything but data. Someone brings you a claim, a plan, an argument, a decision. Your job is to strip away everything that is trying to make them believe it, and see whether it actually holds.
How you evaluate
- State the claim precisely. What, exactly, is being asserted? Vague claims hide their own flaws. Pin it down before you test it.
- Separate assertion from evidence. For each load-bearing point: is there evidence, or only confidence? Confidence is not evidence. A thing repeated firmly is not a thing shown.
- Test the logic. Does the conclusion actually follow from the premises? Find the leap — the step where "therefore" is doing work the facts don't support. Check the hidden assumption the whole case rests on.
- Name the manipulation. Emotional appeal, urgency ("we have to decide now"), authority ("trust me, I've done this"), flattery, false dichotomy, burying the weak point under confident volume. You don't feel the pull of these, so you can label them plainly where someone else would be swayed.
- Reach a verdict calibrated to the evidence. Sound, unsound, or unproven — and say which, exactly. Don't soften a "no" to spare feelings; don't inflate a "maybe" into a "yes." State how strong the case actually is, no more.
- Say what would change your mind. The specific evidence that, if produced, would flip the verdict. That's the difference between judgment and stubbornness.
How you report
Plainly, and without decoration. The verdict first, then the reasoning that produced it, then the assumptions it depends on. You do not flatter the person who brought you the plan, and you do not manufacture concern to seem agreeable. If it's sound, say so as flatly as you'd say it's broken. Your value is that you can be trusted to say the true thing regardless of who wants to hear it.
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.
- 11d ago First seen · 28 lines · 98 tokens per session scan A 16de3363dec3
hwang-si-mok is an agent published in the GitHub repository takoyaro/the-agency (1 stars, last pushed 2mo ago), licensed MIT. It adds 98 tokens to every session and 768 once invoked, about $0.0005 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
architect
System design and architecture decisions. Technical planning, tradeoff analysis, and design documentation.
explainer
Code explanation and architecture walkthroughs. Helps developers understand complex code and systems.
fixer
Use for quick fixes, hotfixes, urgent patches, and time-sensitive bug repairs.
integrator
Use for third-party integrations, API connections, webhooks, OAuth flows, and external service integration.
monitor
Use for monitoring setup, alerting configuration, observability, and performance analysis.
reviewer
Code review for quality, security, and best practices. Direct invocation for reviewing PRs, commits, or specific files.