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 sergeyizmailov/knowledge-delta-skills --skill web-scrapinggit clone --depth 1 https://github.com/sergeyizmailov/knowledge-delta-skillsWrote 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/sergeyizmailov/knowledge-delta-skills/web-scraping)<a href="https://agentmods.dev/skills/sergeyizmailov/knowledge-delta-skills/web-scraping"><img src="https://agentmods.dev/badge/skills/sergeyizmailov/knowledge-delta-skills/web-scraping/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/sergeyizmailov/knowledge-delta-skills/web-scraping"><img src="https://agentmods.dev/badge/skills/sergeyizmailov/knowledge-delta-skills/web-scraping.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 | $0.00085 | $0.01222 |
| Opus 5 | $0.00043 | $0.00611 |
| Sonnet 5 | $0.00017 | $0.00244 |
| Haiku 4.5 | $0.00009 | $0.00122 |
Grade A, and why
web-scraping scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 84 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Mass Scraping — Crawl4AI + Camoufox
Verified 2026-08-27 (macOS arm64). Both engines and all three templates smoke-tested.
Install (if crawl4ai-python is not on PATH)
- crawl4ai —
github.com/unclecode/crawl4ai· PyPIcrawl4ai - camoufox —
github.com/daijro/camoufox· PyPIcamoufox[geoip] - curl_cffi —
github.com/lexiforest/curl_cffi· PyPIcurl_cffi
Dedicated venv, then crawl4ai-setup and python -m camoufox fetch. Four traps, each verified here:
- Python 3.10–3.13 only. 3.14 is not supported — if
python3is 3.14, install 3.13 explicitly. crawl4ai, notcrawl4ai[all]. The extra pulls torch + transformers (GBs) that scraping never uses.crawl4ai-setup, notplaywright install chromium. Setup also installs Patchright, which rung 3 needs; without it the undetected adapter dies.- Shim the venv python with a wrapper, never a symlink. A symlinked venv python resolves
sys.prefixto the system interpreter and cannot import crawl4ai:printf '#!/bin/sh\nexec "$VENV/bin/python" "$@"\n' > ~/.local/bin/crawl4ai-python
Entry points
crwl https://example.com -o md -bc # one page → markdown
crawl4ai-python script.py # anything bigger
Both are shims into the dedicated venv. System python3 cannot import crawl4ai/camoufox — always
crawl4ai-python or the venv binary directly. Don't pip-install these into another environment.
crwl: -o md · -o all (JSON) · -bc (bypass cache) · --deep-crawl bfs --max-pages N.
-o json fails without an extraction schema — use -o all.
Escalation ladder — climb only on failure
| # | Approach | Cost | When |
|---|---|---|---|
| 0 | curl_cffi (TLS/JA3 impersonation, no browser) |
~10ms/page | Try first. Clears more than expected — see 03 |
| 1 | crwl / plain AsyncWebCrawler |
~1s/page | Page needs JS |
| 2 | BrowserConfig(enable_stealth=True) |
~0 | Rung 1 → challenge/403 |
| 3 | UndetectedAdapter() (= Patchright) |
+~1s/page | Rung 2 still blocked |
| 4 | Camoufox + raw:// bridge |
200MB/instance, slow | Chromium fingerprint-blocked |
| 5 | Residential proxies | $ | Blocked by IP, not fingerprint |
What ships with it
6 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.
- 10d ago First seen · 84 lines · 85 tokens per session scan A 085b47a0120b
web-scraping is a skill published in the GitHub repository sergeyizmailov/knowledge-delta-skills (3 stars, last pushed 2d ago), licensed MIT. It adds 85 tokens to every session and 1,222 once invoked, about $0.0004 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 2026-08-31.
Other skills, from other repositories
browserwing-admin
Manage and operate BrowserWing — an intelligent browser automation platform. Install dependencies, configure LLM, create/manage/execute automation scripts, use AI-driven exploration to generate scripts, browse the script marketplace, and troubleshoot issues.
qa
QA test your code changes by reading your git diff, choosing the right validation path for frontend/browser and backend changes, and reporting pass/fail with evidence.
architecture-feature-first
Use when creating a feature, designing folder structure, adding repositories/services/view models, wiring dependency injection, or deciding which layer owns logic.
flutter-use-column-row-first
Use when building any Flutter screen or component to choose responsive Row, Column, Expanded, Flexible, and Spacer layouts before fixed-size or coordinate-based alternatives.
karpathy-llm-wiki
Use when building or maintaining a personal LLM-powered knowledge base. Triggers: ingesting sources into a wiki, querying wiki knowledge, linting wiki quality, 'add to wiki', 'what do I know about', or any mention of 'LLM wiki' or 'Karpathy wiki'.
firebase-ai
Use when setting up firebaseai, generating text/chat with Gemini, streaming AI output, building multimodal prompts, or handling AI errors.