القائمة
Turn a URL into a durable knowledge artifact: searchable markdown summary + local raw text archive, so you can recall it later without revisiting the web.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Use this when you need a complete interactive report as a single HTML file with zero install dependencies — covers charts, tables, diagrams, math, KPIs, and prose out of the box.
Use this as the default toolset for coding sessions when you want faster navigation, search, file inspection, and git workflow execution with lower command overhead: tools: - fzf - ripgrep - bat - delta - lazygit - starship - zoxide - eza - atuin - yazi
Out-of-band codebase exploration using a cheap/fast model. Keeps your context clean while a cheap model runs 30-80 commands in parallel. Tunable precision and output size - works for broad orientation or precise targeted questions.
Deterministically fixes broken OpenAI Deep Research markdown citations without using an LLM: creates a .bkp backup, rewrites citation markers, rebuilds references, and runs strict regex validation.
Explore a codebase to gather context for coding tasks. Preserves context window and offloads work to cheaper models.
Dynamically generate and execute a specialized agent for any task. Use when the user has an arbitrary task that would benefit from a custom-generated agent.
| name | website-scraper |
| description | Turn a URL into a durable knowledge artifact: searchable markdown summary + local raw text archive, so you can recall it later without revisiting the web. |
Archive web pages into a personal knowledge base with structured summaries and preserved source text.
./run.sh "https://example.com/article"
# Archive a blog post
./run.sh "https://simonwillison.net/2024/Dec/19/one-shot-python-tools/"
# Archive a GitHub README page
./run.sh "https://github.com/SWE-agent/mini-swe-agent"
# Use a custom archive location
DATA_DIR=~/research ./run.sh "https://arxiv.org/abs/2501.09891"
Per URL, the scraper writes:
{year}/{date}_{slug}.md — summary + YAML frontmatter{year}/raw/{date}_{slug}.txt — raw extracted text (for LLM context){year}/README.md — auto-updated index rowFetch page → extract clean text (trafilatura) → generate validated summary/frontmatter → save summary + raw text.
Cost: ~0.5 cents/page. Benefit: permanent, searchable, LLM-ready references.