Guide · Agent skill
How to use the Credit Analysis (Fixed Income) agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The credit-analysis skill is a set of written instructions, in a file called SKILL.md, that tells an AI agent how to handle fixed-income and credit questions with a structured framework. Once it is installed, the agent covers bond pricing and yield, rating scales, default-risk models and Chinese city investment platform (LGFV) bonds, and it points to a tested pricing function instead of re-deriving formulas each time. This guide is for readers who have never used agent skills and want to install it, ask it useful questions and understand its limits.
Credit Analysis (Fixed Income)HKUDS/Vibe-Trading · MIT Download the skill Quick start
- 1Download the ZIP. Get the credit-analysis ZIP from the skill's page on this site. The package contains SKILL.md.
- 2Unzip it into a skills folder. Place the folder in ~/.claude/skills/ for all projects, or in .claude/skills/ inside one project.
- 3Or upload it in the Claude apps. Open Settings, go to Capabilities, then Skills, and upload the ZIP.
- 4Set up the Python code. Make sure a Python environment can import src.quantlib.fixedincome from the Vibe-Trading repository, since the pricing functions live there.
- 5Ask a specific question. Give coupon, face value, yield and periods, or the balance-sheet inputs, and ask in plain language. The agent will use the framework and the bundled functions where they apply.
What the credit-analysis skill makes the agent do
A skill is a folder of instructions that an AI agent loads when a request matches its topic. This one covers credit analysis for fixed-income products, with a focus on Chinese bond markets. The skill description in its file reads: fixed-income and credit analysis, covering credit bond ratings, spread analysis, default-risk assessment, LGFV bond research, and convertible bond pricing and strategy.
The framework inside the file covers several areas. These include bond pricing and yield (YTM, duration, convexity, DV01), mapping ratings between S&P, Moody's and Chinese scales, default-risk models (Altman Z-Score, Merton, KMV and scorecards), ABS and MBS cash-flow metrics, and LGFV analysis.
- Bond pricing and yield: YTM, Macaulay and modified duration, convexity, DV01 and key rate duration
- Rating mapping between S&P, Moody's and Chinese rating scales
- Default-risk models: Altman Z-Score, Merton structural model, KMV expected default frequency and credit scorecards
- ABS/MBS metrics such as CPR, SMM and tranche structure
- LGFV (city investment platform) bond analysis, including a four-part assessment model
- Credit spread concepts such as G-Spread, Z-Spread and OAS
How to install the credit-analysis skill
There are two ways to install it. The first is to download the ZIP file from the skill's page on this site and unzip it into your skills folder. For a personal setup, use ~/.claude/skills/. For a single project, use the .claude/skills/ folder inside that project. After unzipping, the folder should contain SKILL.md.
The second way is to upload the ZIP in the Claude apps. Open Settings, go to Capabilities, then Skills, and upload the file there. Either way, the agent can load the skill when a matching question comes up.
- Option A: download the ZIP from this page and unzip it into ~/.claude/skills/ for all your projects
- Option B: unzip it into .claude/skills/ inside one project folder
- Option C: in the Claude apps, go to Settings → Capabilities → Skills and upload the ZIP
How to trigger the skill with natural requests
You do not need a special command. Ask a question in plain language that matches the skill's topics. The file says the skill should be used first for questions about bond pricing and YTM, duration and convexity, corporate credit ratings and default probability, credit spreads, LGFV and ABS/MBS credit, interest-rate risk such as DV01, and the structure of the Chinese fixed-income market.
Give the agent the numbers it needs. The framework works on inputs such as coupon, face value, yield, number of periods, or balance-sheet items for a Z-Score. Requests that include the inputs get more useful answers than vague ones.
- Ask for a price: What is the price of a 5-year 5% annual coupon bond at a 4% yield?
- Ask for risk: How much does the price move if yields rise 10 basis points, using modified duration and DV01?
- Ask for a screen: Calculate the Altman Z-Score for this manufacturing company and tell me which zone it falls in.
- Ask for a rating comparison: Map AA- on the Chinese scale to the S&P and Moody's equivalents.
- Ask for an LGFV view: Walk me through the four-part city investment platform framework for this issuer.
Worked example: pricing a bond with the bundled function
For pricing, the skill tells the agent to import the function from the repository instead of writing the formula out again each time. The function bond_price takes face value, coupon rate, yield, number of periods and payments per year. The file's own example uses a 5-year bond with a 5% coupon and a 4% yield, paid annually, and gives a result of about 104.45.
The function uses discrete compounding by default. It also returns the clean price on a coupon date. If you settle between coupon dates, the file says to add accrued interest from the accrued_interest function, which supports several day-count conventions.
- compounding can be discrete (the default) or continuous
- day_count options in accrued_interest are ACT/365F (default), ACT/360, ACT/ACT, 30/360 and 30E/360
from src.quantlib.fixedincome import bond_price; bond_price(face=100, coupon_rate=0.05, ytm=0.04, n_periods=5, freq=1)What you need and safety notes before running it
The skill needs a Python environment that can run the src.quantlib.fixedincome functions from the wider Vibe-Trading repository. The skill file alone does not contain those functions, so the pricing examples will not run from the ZIP by itself. The listing says the skill runs scripts and does not use the network.
Because it runs scripts, read the code in the repository before you run it on your machine, as you would with any code. The file itself also says the models are simplified, so treat outputs as a starting point for a professional review, not as a credit decision.
- Requires: a Python environment with the Vibe-Trading repository code available (src.quantlib.fixedincome)
- Runs scripts: yes
- Uses network: no, so the skill does not fetch live market data
- Review any script you run, since the skill executes code
Tips and limits to know
The file contains several warnings. It says Chinese credit ratings run high, so an AA rating domestically is roughly comparable to a BBB- internationally, and the outlook should be considered too. It says the Altman Z-Score does not apply to financial institutions and that its parameters need recalibration for China. It also contains no data sources, so bring your own current prices, ratings and financial statements.
Use the skill for learning and screening, and use the convertible-bond skill for the equity option part of convertible bonds. The file says that topic is covered there and this skill focuses on the straight-bond value.
- Supply your own data: the skill has no real-time market feeds
- Check outputs against a professional model or a reviewer before any investment or lending decision
- For convertible bonds, use the separate convertible-bond skill for the conversion option
What people use it for
- Practising bond math for a finance course, such as pricing a bond and checking how its price moves with yield using duration and convexity.
- Screening Chinese fixed-income issues with a formula-based view of ratings, default risk and LGFV factors.
- Building or extending a fixed-income tool on the Vibe-Trading codebase, using the bundled pricing functions as a base.
- Explaining hedging to a team by calculating DV01 and a hedge ratio for a position.
Frequently asked questions
How do I install the credit-analysis agent skill?
Download the ZIP from this page and unzip it into ~/.claude/skills/ for all your projects, or into .claude/skills/ inside one project. You can also upload the ZIP in the Claude apps under Settings, Capabilities, then Skills.
Is the credit-analysis skill free to use?
The listing states the licence is MIT. Check the licence file in the repository for the full terms before you reuse or redistribute the code.
What do I need to run the credit-analysis skill?
You need a Python environment that can run the src.quantlib.fixedincome functions from the wider Vibe-Trading repository. The skill file alone does not include those functions. The skill runs scripts and does not use the network.
Does the skill give live bond prices or ratings?
No. The file contains no data sources, so the skill does not provide real-time market data. You need to supply current prices, ratings and financial statements yourself.
Can I use the Altman Z-Score for banks?
No. The skill file says the Altman Z-Score does not apply to financial institutions because their leverage is defined differently. It also says Chinese market parameters need recalibration.
How does this compare with building the formulas myself?
The skill sends the agent to a tested pricing function in the repository instead of rewriting formulas each time. For learning the formulas, reading the SKILL.md framework alongside the code is useful. For production decisions, use a professional model and a reviewer.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use this skill if you are a finance student, an analyst screening Chinese fixed-income issues, or a developer building on the Vibe-Trading codebase, and you want formula-based answers with clear assumptions. Pick something else if you need live market data, a licensed rating service, or a model you can rely on for lending or investment decisions without professional review.
Download the skill