Url ingest

Skill ternary-ai/skills/skills/basic/url-ingest

Fetch and analyse any public URL — news article, press release, SEC filing, earnings transcript, research note. Auto-triggered whenever the user supplies an https:// or http:// link.From its SKILL.md

Install

npx -y skills add ternary-ai/skills --skill url-ingest

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

3.4 KB, 766 tokens by cl100k_base, as published. Nobody here has run it

URL Ingest

Purpose: Retrieve the full text of a public URL the user provides and deliver a substantive analysis of its content, with investment framing when relevant.

Trigger: User message contains an https:// or http:// URL, regardless of any other text in the message. This skill takes absolute priority — do not wait for a skill dispatch prefix.

Process

Extract the URL — Pull the bare URL from the user message. If multiple URLs are present, process each in sequence.
Fetch the content — Call fetch_url(url, max_chars=8000). Use max_chars=8000 as the default — articles need room. If the response contains "paywall", "subscribe", "sign in", or fewer than 200 words of body text, note it clearly and summarise from what is available.
Classify the content type — Determine what was fetched:
- News article / blog post
- Press release / investor relations page
- SEC filing (10-K, 10-Q, 8-K, DEF 14A, etc.)
- Earnings call transcript
- Research note / analyst report
- Other (describe)
Analyse — Deliver a structured summary appropriate to the content type:

News / Press Release:
- One-line recap: source, date, core claim.
- Key facts (3–6 bullets).
- What this means for the company / sector.
SEC Filing:
- Filing type, period, date.
- Material disclosures: revenue, guidance, risks, management commentary.
- What changed vs. prior filing (if context available).
Earnings Transcript:
- EPS beat/miss, revenue beat/miss.
- Management tone: confident / cautious / defensive.
- Key forward guidance quotes (verbatim short snippets).
- Analyst Q&A highlights.
Research Note:
- Analyst firm, rating, price target.
- Core thesis in 2 sentences.
- Key estimates and assumptions.
Investment Framing — For any content that is material to the active ticker (or to a ticker named in the article), append:
Investor Angle
- Bull case implication: …
- Bear case implication: …
- What to watch next: …
Skip this section if the content is clearly unrelated to investable assets.
Thesis update — If the article is material (earnings, deal, restructuring, regulatory action, management change, guidance revision), populate the thesis JSON field with a brief new section appended to the existing thesis. For routine news or background reading, leave thesis empty.

Fallbacks

Paywall / login wall → summarise from any visible text + headline; state clearly that the full article was behind a paywall.
Fetch error → report the exact error in chat; do not hallucinate content.
Non-English content → translate key passages; flag language in the recap.
PDF / binary → fetch_url will return raw text extraction; parse what is readable.

Output Rules

Lead with the source domain and content type (e.g. "Yahoo Finance · News article").
Do NOT write "I fetched..." — present content directly.
Cite publication date and author when visible in the fetched text.
Keep summaries tight: short paragraphs, bullet lists for key facts.
Never fabricate quotes — only use text that appears in the fetched content.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.