Ken-Technology/cold-email-skills / knowledge-setup
knowledge-setup
VerifiedSet up a company knowledge folder as plain markdown files: company overview, ideal customer profile, pricing, proof, competitors, voice and instructions, each fact with its source. Use when someone wants their agent to know their company before writing outbound. Triggers on "set up my company knowledge", "write our company files", "capture our voice".
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 8.4 KB |
| references/website-scraping.md | markdown | 3.4 KB |
references/website-scraping.md
View on GitHub# Website Scraping Skill
Pick the pages of a company website that explain what it sells and to whom, and read them as markdown with your own web tools (a page fetch or a browser), or from page text the user pastes. The pages feed client research, the company knowledge folder, campaign strategy, and copy.
## Tools
Use the web tools your agent has: one that fetches a page as markdown or text and, when there is one, a site map or site search. To map a site without one, read the homepage and collect its links, and `/sitemap.xml` when the site has one.
Read one URL per call. For 8 pages it is the simplest and the most reliable path.
With no web tool, ask the user for the pages' text or the links they want read, and say that you cannot read the site yourself.
## The site
Read the site the user names. When they name none, use the website in `research.md` or the knowledge folder's `company-overview.md`. With no website anywhere, ask for it.
## Page Selection
If the user gives explicit URLs, scrape exactly those: no filtering, no cap.
Otherwise map the site and pick **at most 8 pages**:
1. **Same host only** - the site's own domain (a leading `www.` does not matter). Skip subdomains such as `blog.`, `docs.`, `app.`, `help.`, and every external link.
2. **Priority order** - homepage; about (`/about`, `/company`, `/team`, `/our-story`); product (`/product`, `/products`, `/solutions`, `/services`, `/platform`, `/features`, `/how-it-works`, `/process`); pricing (`/pricing`, `/plans`, `/packages`); customers (`/customers`, `/case-studies`, `/success-stories`, `/testimonials`, `/reviews`); industries and use cases (`/industries`, `/use-cases`, `/verticals`); integrations (`/integrations`, `/partners`); why-us.
3. **Skip** - blog posts and articles (`/blog/*`, `/news/*`, `/press/*`, single resource items), careers and jobs, legal (`/privacy`, `/terms`, `/legal`, `/cookie-policy`, `/gdpr`), login, signup, account, and app pages, search results and pagination (`?q=`, `?page=`, `/page/2`), media files (`.pdf`, `.jpg`, `.png`, `.mp4`), `/admin`, `/api`, `/webhook`, single event or webinar pages.
4. **One URL per page** - drop query strings, fragments, and trailing slashes before comparing, so the same page is never scraped twice.
If map fails or returns nothing usable, scrape the homepage alone.
## Scraping
- A page that fails is skipped; keep going. Never abort a site for one failure.
- Site unreachable: report it and stop for that site.
- Never echo scraped markdown back into the conversation. Report one line per page: URL and `ok` or `failed`, then how many URLs the map found and how many pages were read.
- With several sites, finish one before starting the next.
## Save pages
When the pages should be kept for later skills, save each one in the campaign workspace (`./cold-email/{slug}/`) right after reading it, before the next page:
- Name: `research/website/<page-slug>.md`. `/` -> `research/website/index.md`, `/about` -> `research/website/about.md`, `/products/x` -> `research/website/products--x.md`. Nested path segments join with `--`.
- Content: front matter with `url`, `title` (from the page, or empty), and `read_at` (ISO timestamp), then the page markdown.
- One file per page: when the file exists, overwrite it. Never create duplicates.
**client-research** reads the saved pages from there. For **knowledge-setup**, saving is optional: each fact goes into its knowledge file with the page and the date as its source.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |