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 MadaraUchiha-314/the-loop --skill writinggit clone --depth 1 https://github.com/MadaraUchiha-314/the-loopWrote 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/madarauchiha-314/the-loop/writing)<a href="https://agentmods.dev/skills/madarauchiha-314/the-loop/writing"><img src="https://agentmods.dev/badge/skills/madarauchiha-314/the-loop/writing.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 | $0.00112 | $0.00974 |
| Opus 5 | $0.00056 | $0.00487 |
| Sonnet 5 | $0.00022 | $0.00195 |
| Haiku 4.5 | $0.00011 | $0.00097 |
Grade A, and why
writing 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.
How it starts
The opening of the file, as written. The whole thing — 86 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Writing for a human reader
A reviewer's job is to say yes or no. Every word between them and that decision is a cost you are charging them. Write to be approved, not to be admired.
The rule: say it once, to a named reader, in the fewest words that survive review.
Who is reading
Name the reader before the first sentence. A reviewer of design.md knows the codebase
and wants the shape of the change. A capability.md reader wants current behaviour and
does not care how it arrived. A ticket comment is read on a phone. The same fact is a
paragraph for one and a table row for another.
The spine
Every explanatory document answers four questions, in this order:
- What was broken — the situation, and why it mattered enough to open a ticket.
- What we did about it — the resolution, stated as a decision, not a tour.
- What it costs — the trade-off taken, the thing given up, the risk carried.
- What to check — where the reader should look first, and what would falsify it.
Front-load each one. A section's first sentence carries its conclusion; the rest is support the reader may skip. Read only the first sentence of every section — if that alone tells the story, the document is shaped right.
Length follows the change
There is no word limit, and there should not be. A two-line bug fix and a new
subsystem are both design.md, and a number that fits one is wrong for the other. A cap
would only teach you to move prose into an appendix.
The test is not length, it is density: can any sentence come out without losing information? While the answer is yes, the document is too long — at 200 words or 2000. Three cuts nearly always find the slack: the sentence restating the heading, the sentence restating the sentence before it, and the adjective a number would replace.
Concision is about words, not coverage. A gated section stays even when it is empty;
say so in one sentence, and why. Deleting ## Security considerations to shorten a
document is fraud, not editing.
What ships with it
1 file 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.
- 8d ago First seen · 86 lines · 112 tokens per session scan A 7a20d51b9669
writing is a skill published in the GitHub repository MadaraUchiha-314/the-loop (4 stars, last pushed 4d ago), licensed MIT. It adds 112 tokens to every session and 974 once invoked, about $0.0006 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 skills, from other repositories
offer-clarifier
Force clarity on what you sell, who pays, and why now. Use before working on ICP, positioning, pricing, or outreach - the downstream work is garbage if the offer is vague.
sequence-architect
Design a 4-6 touch outreach cadence with channels, timing, tonal arc, and branching. Use when starting a new sequence from scratch given an ICP and offer.
gtm-motion-picker
Decide between PLG, sales-led, hybrid, or community-led GTM motion based on ACV, time-to-value, buyer complexity, and self-serve readiness. Use when launching or repositioning a product, or when sales costs are out of whack with ACV.
icebreaker
Draft 3 distinct opening messages for a specific prospect, anchored on a why-now and a why-this-person. Use when sending a high-leverage first touch where personalization is worth the time.
pricing-teardown
Run a competitive pricing teardown and produce a pricing recommendation. Pulls 5-10 competitor pricing pages, normalizes to a common axis, identifies packaging anti-patterns, and runs a Van Westendorp sanity check on the user's own price.
icp-definer
Define an Ideal Customer Profile from observed paying-customer data, churn data, and qualitative signals. Use when defining or tightening ICP, when sales efficiency drops (rising CAC, falling close rate), or when support load suggests wrong-fit customers.