Card compare

Skill jiahongc/credit-card-skills/.claude/skills/card-compare

Agent Skills suite for researching US credit cards — works with Claude Code, Codex, Cursor, Gemini CLI, and any Agent Skills-compatible tool

Install

npx -y skills add jiahongc/credit-card-skills --skill card-compare

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.
5 stars5 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.

What its author says it does

Copied from the file, not written here

Return a side-by-side comparison of two major-US credit cards across fees, earning rates, credits, transfer partners, and key benefits. Use when the user is deciding between two cards.

SKILL.md

2.6 KB, 596 tokens by cl100k_base, as published. Nobody here has run it

Card Compare

Goal

Return a compact side-by-side comparison of two exact card variants.

Search Strategy: search-required

Fetch both issuer pages in parallel + up to 5 secondary sources total for current offer details. Prefer NerdWallet, The Points Guy, Doctor of Credit, One Mile at a Time, and Bankrate.

Workflow

Parse two card names from the input (separated by "vs", "versus", "or", or a comma).
Resolve each card using ../card-identity/SKILL.md. If either is ambiguous, return a numbered choice list for the ambiguous card and stop.

Run one Brave Search API call (see search_method in ../card-shared/source-policy.yaml):

curl -sS "https://api.search.brave.com/res/v1/web/search?q=CARD_A+vs+CARD_B+compare&count=10" -H "X-Subscription-Token: $BRAVE_API_KEY"

If $BRAVE_API_KEY is not set, fall back to WebSearch.

Fetch pages — pick the issuer URL for each card and up to 2 secondary URLs (prefer nerdwallet.com and thepointsguy.com) from the search results. Fetch in parallel:
```
curl -sS -L "URL" | sed 's/<[^>]*>//g' | tr -s '\n' | head -200
```
Search snippets are too shallow for accurate comparisons — the actual pages have complete rate tables, credit lists, and benefit details.
Apply confidence handling from ../card-shared/confidence-rules.md.
Return compact markdown using the card-compare contract in ../card-shared/command-contracts.yaml.
YAML is internal only — do not include it in user-facing output.
Do not show a "Why It Matters" section.

Required Sections

## 💰 Fees
## 📈 Earning Rates
## 🏷️ Credits
## 🔄 Transfer Partners
## ✈️ Key Benefits
## 🏆 Bottom Line
## 📋 Confidence Notes
## 🔗 Sources — numbered list of URLs fetched, as markdown hyperlinks with short "Site - Topic" labels, e.g. [NerdWallet - CSR vs Amex Gold](https://...)

Omit ## Card Identity when both matches are confident. Use a two-column format (Card A | Card B) for the comparison sections. The Bottom Line section should give a concise factual summary of which card wins in each major dimension — not a recommendation.

What ships with it

Read from the repository

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