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 skills/friedbotstudio/baseline/technical-writingnpx skills add friedbotstudio/baseline --skill technical-writinggit clone --depth 1 https://github.com/friedbotstudio/baselineWrote 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/friedbotstudio/baseline/technical-writing)<a href="https://agentmods.dev/skills/friedbotstudio/baseline/technical-writing"><img src="https://agentmods.dev/badge/skills/friedbotstudio/baseline/technical-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 | $0.00094 | $0.02889 |
| Opus 5 | $0.00047 | $0.01444 |
| Sonnet 5 | $0.00019 | $0.00578 |
| Haiku 4.5 | $0.00009 | $0.00289 |
Grade A, and why
technical-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 today.
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 — 183 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Documentation that reads as machine-written is rarely wrong and rarely ugly. It fails on range. Every sentence lands near the same length, every paragraph opens with the same rhetorical move, every claim arrives at the same pitch of certainty. A reader cannot say which sentence is wrong, only that the whole thing feels extruded.
This skill is built from measurement rather than taste. A corpus of 28 documentation pages published before 2022 — 113,887 words of body prose from SQLite, PostgreSQL 12, Python 3.8, Django 3.2, the Rust book, Effective Go, Pro Git, nginx, Backbone, 12factor and Redis — was parsed and profiled on 24 axes. references/corpus-profile.md carries the full table. Every number below comes from it.
The finding that matters most. The corpus writes longer, more qualified, more conditional prose than a de-slopped draft does. Across the shipped pages this skill was written to fix, not one sentence in 1,262 ran past 30 words; professional documentation runs 11.5% of its sentences past 30 words. Stripping complexity is what made those pages read as generated. Bans alone cannot fix that, because a ban can only remove.
The gate
node .claude/skills/technical-writing/measure.mjs --type <reference|explanation|tutorial|howto> <file>
Accepts HTML, Markdown, or a rendered page. It returns a weighted deviation score against the corpus bands and names each axis that is out of range, worst first. Threshold 4.5. For scale: the corpus median is 0.14, the corpus 90th percentile is 4.93, and the site pages this skill was built to repair scored 5.83 to 30.2.
A page is not finished until this exits 0. The score is calibrated, not decorative — at 4.5 it passes 86% of the professional corpus and none of the pre-rewrite pages.
Step 1 — Classify. Exactly one type.
The type decides the numeric targets, the heading grammar and the person. A page that mixes types fails on every axis at once, because it is averaging two profiles.
What ships with it
3 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.
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.
- today First seen · 183 lines · 94 tokens per session scan A 2797a0887491
technical-writing is a skill published in the GitHub repository friedbotstudio/baseline (14 stars, last pushed yesterday), licensed Apache-2.0. It adds 94 tokens to every session and 2,889 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 2026-09-03.
Other skills, from other repositories
probe
Report this absolute source path, then read the nested routine.
dev-standards
Enforces development workflows, quality gates, coding standards, and release processes for the deterministic-agent-control-protocol project. Use when implementing features, fixing bugs, refactoring architecture, adding integrations, updating policies, writing tests, updating documentation, or preparing releases.
brainstorming
The Socratic spec-refinement front of /feature, and the planning front of /sprint. Routed to BEFORE any code — it takes a one-line idea and drives it to an approved, concrete spec with testable acceptance criteria. Five gated phases — frame, shape, refine, write, review-and-approve. No implementation and no handoff to…
commit-gate
The only path to a commit. Routed to when the user invokes /commit or otherwise instructs codeArbiter to persist staged changes. Nine gated phases — permission, branch, classification, verification (test/lint/secrets), behavioral proof, diff review, selective stage, message, commit. Nothing reaches version control…
context-check
Optional manual drift audit — report stale provenance-tracked docs (via provenancelib drift detection across .codearbiter/.provenance/), then per stale doc offer re-scout / re-baseline / defer. Not the daily loop; commit-gate auto-heal owns routine maintenance.
context-creation
The brownfield back-fill. Routed to by /create-context, and by startup when .codearbiter/CONTEXT.md lacks the body marker but source code exists. Six gated phases — pre-flight, scout dispatch, synthesis, gap interview, write, lock. Reads the existing codebase through parallel scouts, drafts every surviving…