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/webjsdev/webjs/webjs-blog-writenpx skills add webjsdev/webjs --skill webjs-blog-writegit clone --depth 1 https://github.com/webjsdev/webjsWhat 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 | $0.00133 | $0.03377 |
| Opus 5 | $0.00067 | $0.01688 |
| Sonnet 5 | $0.00027 | $0.00675 |
| Haiku 4.5 | $0.00013 | $0.00338 |
Grade A, and why
webjs-blog-write scanned grade A with 1 finding 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 yesterday.
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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
3. **For a feature post, verify every factual claim by dogfooding before you publish.** This is not optional for a post that describes runtime behaviour. Scaffold a real app (`webjs create`) and run the verification's in How it starts
The opening of the file, as written. The whole thing — 104 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Write a webjs blog post
The repo publishes engineering blog posts from the repo-root blog/ directory. The website reads them directly (website/modules/blog/queries/list-posts.server.ts, BLOG_DIR = <repo>/blog), parses the front matter, and sorts by date descending. A post is one markdown file at blog/<slug>.md. This skill is how a new post lands in the author's voice, ranks for search, and states only claims that are actually true, without a later cleanup pass.
Step 0 (mandatory, do this FIRST): analyze the author's current style
Do NOT write from a fixed style spec. The voice is defined by the corpus, and the corpus changes over time, so a hard-coded spec goes stale. Before writing a single sentence:
- Read a broad sample of
blog/*.md(at least six to eight posts spanning different topics and dates, including the two or three most recent bydate). Read whole posts, not just the openings. - From that reading, note for yourself: how the author opens a post, how paragraphs are paced, how code is introduced, how sections are titled, how a post closes, and which words and rhythms recur. The sections below capture what has been true so far, but the corpus you just read is the authority. If it has drifted from this description, follow the corpus and update this skill.
The patterns that have held so far (confirm them against what you read):
- First person, opinionated, honest. The author writes as themselves ("I wanted the opposite", "kept biting me", "I did not find one I enjoyed using"). Personal origin and motivation are welcome.
- A concrete hook, never a definition. Posts open on a scenario, an attack, a painful task, or a plain-language framing, not on "X is a...". A handful of posts use a signature "Let me start with..." or "Let me show you" opener; keep that phrasing where the author reaches for it, but the dependable pattern is the concrete hook itself, not the specific words.
- Beginner-friendly. Every acronym or jargon term is expanded in parentheses on first use, for example "SSR (server-side rendering)" or "an async generator (a function that yields values over time instead of returning once)".
- Second person in walkthroughs. Once into the mechanics, the author addresses "you" and shows real
ts/shcode blocks. - The painful old way, then the WebJs way. Most feature posts show the tedious status-quo pattern first (a useEffect + spinner, thirty lines of optimistic bookkeeping, a native-addon compile), then the WebJs version that removes it.
- Section headers are single
#, sentence case, often with a trailing clause. For example "# The classic answer, and why it is fiddly" or "# What the sourcemap layer cost". - Length runs substantial, commonly around 1000 to 1600 words, with deeper posts closer to 2200. Match the length of comparable posts you read in Step 0. Substance over padding, and never pad to hit a number.
- Vary the closing. Do NOT reach for a generic "# The takeaway" header. A closing header that repeats verbatim across the blog is the clearest formulaic tell there is, no matter how good each post is on its own. Most posts should just land their point in the final paragraph with no labeled closing section at all. When a post does want a labeled close, give it a header specific to that post's argument, and check that no other post already uses it. Treat any header that shows up on more than one or two posts as a smell.
- Contrast with the tool the reader knows. React, Next.js, Rails, Lit, Remix, Astro. Explain what that tool does and why WebJs chose differently, on WebJs's own terms.
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.
- yesterday First seen · 104 lines · 133 tokens per session scan A a3d0c30723e1
webjs-blog-write is a skill published in the GitHub repository webjsdev/webjs (109 stars, last pushed 10d ago), licensed MIT. It adds 133 tokens to every session and 3,377 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
javascriptcore-garbage-collector
JSC GC reference for Bun. Use for use-after-free, JS object leaks, "collected too early", or when touching WriteBarrier, visitChildren, visitAdditionalChildren, JSRef, JSC::Strong/Weak, hasPendingActivity, ensureStillAlive, addOpaqueRoot, reportExtraMemoryAllocated, IsoSubspace, HeapAnalyzer, finalize.
implementing-jsc-classes-cpp
Implements JavaScript classes in C++ using JavaScriptCore. Use when creating new JS classes with C++ bindings, prototypes, or constructors.
implementing-jsc-classes-rust
Creates JavaScript classes using Bun's Rust bindings generator (.classes.ts). Use when implementing new JS APIs in Rust with JSC integration, prototypes, or constructors.
rust-system-calls
Guides using bunsys for system calls and file I/O in Rust. Use when implementing file operations, opening fds, or any syscall path instead of std::fs or libc.
slowest-tests
Find the top-N slowest test files in CI from a recent BuildKite run, optionally posting the results to a Slack channel as a formatted table. Use when asked to find slow CI tests, "what's making CI slow", or to post a slow-test report to Slack.
writing-bundler-tests
Guides writing bundler tests using itBundled/expectBundled in test/bundler/. Use when creating or modifying bundler, transpiler, or code transformation tests.