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/contextgeneric/cgp/dual-reader-prosenpx skills add contextgeneric/cgp --skill dual-reader-prosegit clone --depth 1 https://github.com/contextgeneric/cgpWhat 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 | $0.00124 | $0.02138 |
| Opus 5 | $0.00062 | $0.01069 |
| Sonnet 5 | $0.00025 | $0.00428 |
| Haiku 4.5 | $0.00012 | $0.00214 |
Grade A, and why
dual-reader-prose 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 3d 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 — 208 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Dual-Reader Prose
Good writing serves two kinds of readers at once. Some people scan first and read only what they need. Others read every word in order and want the full picture. Most formats help one group and hurt the other. Bullet points help scanners but frustrate deep readers. Dense prose helps deep readers but loses scanners. This skill shows you how to write for both. Use prose by default, and use bullets only when the content is truly a list.
Write in Plain English
Plain words and short sentences help both readers. A scanner grasps the point faster. A deep reader follows the argument with less effort. So the first rule is simple: say things plainly.
Prefer simple words. Choose the common word over the fancy one. Write "use" instead of "utilize", "show" instead of "demonstrate", and "help" instead of "facilitate". A plain word is easier to read and never sounds thin.
Write short sentences. One sentence should carry one idea. When a sentence runs long or holds two or three ideas, split it. Short sentences are easier to scan and easier to follow.
Cut words you do not need. Remove filler like "in order to", "it is important to note that", and "the fact that". If a word adds nothing, delete it.
Say it directly. Prefer the active voice and a clear subject. Write "the reader stops here" instead of "it is at this point that the reader may cease".
Plain English is not the same as shallow writing. You keep every idea. You just say each one in the clearest way you can.
The Two Reader Types
Readers behave differently because they process information differently. The difference is not random. It reflects how each person reads and remembers.
Scanner-first readers come to a text with a goal. They want to find what they need, skip what they know, and move on. They read the first sentence of a paragraph to decide whether to read the rest. Lists and headers are their map. When a text is all prose with no visual anchors, they feel the writer is wasting their time and making them dig. Bullet points feel fast and respectful. These readers often work at speed, read a lot each day, and have learned to scan on purpose.
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.
- 3d ago First seen · 208 lines · 124 tokens per session scan A 993ab8a797a5
dual-reader-prose is a skill published in the GitHub repository contextgeneric/cgp (248 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 124 tokens to every session and 2,138 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 2026-08-30.
Other skills, from other repositories
effect-patterns-building-apis
Effect-TS patterns for Building Apis. Use when working with building apis in Effect-TS applications.
effect-patterns-concurrency
Effect-TS patterns for Concurrency. Use when working with concurrency in Effect-TS applications.
effect-patterns-building-data-pipelines
Effect-TS patterns for Building Data Pipelines. Use when working with building data pipelines in Effect-TS applications.
effect-patterns-error-management
Effect-TS patterns for Error Management. Use when working with error management in Effect-TS applications.
effect-patterns-making-http-requests
Effect-TS patterns for Making Http Requests. Use when working with making http requests in Effect-TS applications.
effect-patterns-observability
Effect-TS patterns for Observability. Use when working with observability in Effect-TS applications.