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 EpisodeYu/HarnessSmith --skill web-readinggit clone --depth 1 https://github.com/EpisodeYu/HarnessSmithWrote 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/episodeyu/harnesssmith/web-reading)<a href="https://agentmods.dev/skills/episodeyu/harnesssmith/web-reading"><img src="https://agentmods.dev/badge/skills/episodeyu/harnesssmith/web-reading/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/skills/episodeyu/harnesssmith/web-reading"><img src="https://agentmods.dev/badge/skills/episodeyu/harnesssmith/web-reading.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 | $0.00082 | $0.00515 |
| Opus 5.5 | $0.00033 | $0.00206 |
| Sonnet 5.5 | $0.00016 | $0.00103 |
| Haiku 4.5 | $0.00008 | $0.00052 |
Grade A, and why
web-reading 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 2mo 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.
What it actually says
Web reading: get clean text from complex pages
Plain HTTP fetch tools (e.g. the fetch MCP tool) download the raw HTML and
strip tags. That fails on modern sites whose content is rendered by JavaScript
(SPAs, infinite-scroll, app-shell pages): the raw HTML is an empty shell, so
you get little or no real text.
Jina Reader (https://r.jina.ai/) renders the page in a real browser and
returns clean Markdown of the main content. It is keyless (anonymous, rate
limited) — no API key or config needed — so it works in any generated harness
that has a URL-fetching tool.
When to use this
Use Jina Reader whenever a normal fetch of a page returns any of:
- empty or near-empty body, or only nav/menu/footer boilerplate
- "Please enable JavaScript", a loading spinner, or an obvious app shell
- garbled / truncated content that doesn't match what the page clearly contains
- a long, content-heavy article you want as clean Markdown
How to use it
Prepend https://r.jina.ai/ to the original URL and fetch THAT with your
existing fetch / HTTP-GET tool:
https://r.jina.ai/https://example.com/some/article
For example, to read https://news.site/article/123, fetch
https://r.jina.ai/https://news.site/article/123. The result is rendered
Markdown of the page's main content.
Notes:
- Keep the original URL's scheme (
https://...) after the prefix. - It is rate limited when keyless; if you hit a limit or a page is anti-bot protected, say so and fall back to the plain fetch result rather than looping.
- If this harness has the
jina-readerMCP server enabled (with aJINA_API_KEY), prefer itsread_urltool instead — same engine, higher quota, more reliable.
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.
- 2mo ago First seen · 47 lines · 82 tokens per session scan A 5b73288861a4
web-reading is a skill published in the GitHub repository EpisodeYu/HarnessSmith (21 stars, last pushed 3mo ago), licensed MIT. It adds 82 tokens to every session and 515 once invoked, about $0.0003 per session on Opus 5.5. 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
generate
Turn a codebase, an OpenAPI/gRPC/GraphQL spec, or a plain-language description of an API into a production-ready MCP server, using mcp-anything. Use when the user wants to expose software, an API, a CLI tool, or a service as MCP tools for Claude or other agents.
write-terva-skill
Author a new terva skill, a reusable SKILL.md instruction set that terva loads on demand. Install it where terva discovers it, in the project .terva/skills/ or the global $TERVAHOME/skills/. Use for create/add/write-a-skill requests, not extensions.
write-terva-locale
Translate the terva interface into another language, complete or fix a locale, or wrap new strings for translation. Use when the user asks to localize terva or to add or finish a language (fi, pt-BR, ...). Also use to edit translated text or to contribute a translation upstream.
write-terva-lore
Author terva lore. Lore is context that a person authors in files. terva injects it only when it is keyword-relevant, or always-on. Use to create/add/write lore, "world info", a lorebook entry, keyed background, or a book of facts for a character.
handoff
Compact the current session into a handoff document. A fresh session or another agent can pick it up with no need to re-derive context. Use when asked to hand off, hand over, wrap up for another session or agent, or park work mid-stream.
house-style
Apply before you write or edit prose, such as a README.md, documentation, a comment, or a commit message. It covers release notes and skill or persona bodies too. Also use when the user asks you to clean up prose, or to cut machine tells.