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 xberg-io/crawlberg --skill crawlberg-headless-fallbackgit clone --depth 1 https://github.com/xberg-io/crawlbergWrote 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/xberg-io/crawlberg/crawlberg-headless-fallback)<a href="https://agentmods.dev/skills/xberg-io/crawlberg/crawlberg-headless-fallback"><img src="https://agentmods.dev/badge/skills/xberg-io/crawlberg/crawlberg-headless-fallback/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/xberg-io/crawlberg/crawlberg-headless-fallback"><img src="https://agentmods.dev/badge/skills/xberg-io/crawlberg/crawlberg-headless-fallback.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 | $0.00052 | $0.01157 |
| Opus 5.5 | $0.00021 | $0.00463 |
| Sonnet 5.5 | $0.00010 | $0.00231 |
| Haiku 4.5 | $0.00005 | $0.00116 |
Grade A, and why
crawlberg-headless-fallback 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 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.
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 — 145 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Headless fallback
Some pages are unscrapable without a real browser — SPA shells, infinite scroll, Cloudflare interstitials, JS-rendered article bodies. Crawlberg ships with an optional headless-Chrome backend driven by chromiumoxide.
Modes
--browser-mode auto # default — try static first, fall back to browser on JS/WAF
--browser-mode always # skip static, go straight to browser
--browser-mode never # static only, fail closed
auto (default)
The engine fetches statically, then inspects the response. It launches headless Chrome and re-fetches when it sees:
- WAF responses from one of 8 detected vendor fingerprints (Cloudflare, Akamai, AWS WAF, Imperva, DataDome, PerimeterX, F5, plus a generic catch-all).
- SPA shells:
<noscript>warnings, near-empty<body>with heavy JS. - Heuristic JS-render-required signals.
This is the right default. The browser only spins up when needed.
always
Skip the static probe entirely. Use when:
- The user already told you the page needs JS.
- You are scraping a site you know is React/Vue/Svelte SPA.
- You need
<script>-emitted state that never lands in static HTML.
crawlberg scrape https://spa.example.com --browser-mode always --format markdown
never
Static only — the browser path is disabled. Use when:
- You are in a hot loop where a stray Chrome launch would blow the budget.
- You are running in a sandbox without a Chrome binary.
- The user explicitly wants only static fetches.
In never mode, JS-only pages return empty/stub content. Inspect
markdown.content and markdown.warnings before treating the result as
final.
Symptoms that point to headless
In --browser-mode never or when you suspect the auto detector missed a
signal:
markdown.contentis short, nav-only, or just a loading message.status_codeis 200 butmetadata.headingsis empty on a page that clearly has headings.markdown.warningsmentions JS-render-required or WAF detection.- 403/406/503 with WAF response headers (
server: cloudflare,cf-mitigated,x-amz-cf-id,set-cookie: __cf_bm=…).
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.
- yesterday First seen · 145 lines · 52 tokens per session scan A 07b9a8a01ff6
crawlberg-headless-fallback is a skill published in the GitHub repository xberg-io/crawlberg (185 stars, last pushed today), licensed MIT. It adds 52 tokens to every session and 1,157 once invoked, about $0.0002 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 2026-10-15.
Other skills, from other repositories
tree-sitter-language-pack
Parse and extract code intelligence from 372 programming languages using tree-sitter grammars. Use when writing code that parses source, extracts structure/imports/exports/symbols/docstrings/comments, detects a language, runs syntax diagnostics, or produces syntax-aware chunks for LLMs — in Rust, Python…
tree-sitter-language-pack-extracting-code-structure
Use when the user wants structured code metadata from a source file — functions, classes, imports, exports, symbols, docstrings, comments, or syntax diagnostics. Covers ts-pack process feature flags, the JSON result shape, and the default feature set.
tree-sitter-language-pack-detecting-languages
Use when the user wants to know which programming language a file or snippet is. Covers implicit detection in ts-pack parse/process, confirming support with ts-pack list/info, and the SDK detection functions for path, extension, and raw content.
tree-sitter-language-pack-managing-parsers
Use when the user needs to manage the tree-sitter parser cache — prefetch parsers for offline or CI runs, list what is downloaded, inspect a language, find the cache directory, or clean it. Covers ts-pack download, list, info, cache-dir, clean, and init.
tree-sitter-language-pack-using-the-mcp-server
Use when parsing source, extracting code structure, or detecting a language through the tree-sitter-language-pack MCP server's tools, rather than shelling out to the ts-pack CLI. Covers the tool surface, the auto-installing launcher, and when MCP beats the CLI or SDK.
tree-sitter-language-pack-chunking-for-llms
Use when the user wants to split source code into chunks for an LLM context window without breaking syntax mid-construct. Covers ts-pack process --chunk-size, why syntax-aware splits beat fixed-byte splits, picking a size, and the chunk JSON shape.