MCP Tools and Prompts - Credit Benchmark Docs

Your assistant chooses these tools for you based on your question, so you do not need to call them by name. This page explains what each one does, which helps when you want to steer an answer or understand where a figure came from.

Tools

Tool	What it does
Find entities
`cb_match_entities`	Matches company or issuer names to `CB_ID`s, returning ranked candidates. Country and industry hints help with ambiguous names. Up to 50 names per call.
`cb_match_leis`	Matches Legal Entity Identifiers (LEIs) to `CB_ID`s exactly. Up to 50 LEIs per call.
Single entities
`cb_view_entity`	Displays a credit memo exhibit for one entity: the current consensus, rating history with the contributor range, the industry aggregate over the same window, and your bank’s own rating where you contribute.
`cb_get_entity_data`	Returns row-level ratings, PDs, and attributes for a list of entities or a scope, as a snapshot or a time series.
Portfolios and sectors
`cb_get_credit_breakdown`	Summarises credit quality for a portfolio or scope, grouped by investment grade and high yield, sector, country, or another column, including upgrade and downgrade counts.
`cb_get_aggregate_trend`	Returns an aggregate PD and rating trend over time for a scope, optionally split by a column such as `CB_Sector`.
`cb_get_rating_distribution`	Returns the mix of ratings in a scope, at one date or for each publication date over a window.
`cb_get_entity_rating_changes`	Compares each entity’s current rating and PD with a prior date to find upgrades, downgrades, and the largest moves.
Reference
`cb_get_metadata`	Looks up field names, valid industry and geography values, rating scales, and the latest publication dates.
`cb_search_documentation`	Searches this documentation site and returns cited excerpts for methodology questions.

Most data and analytics tools accept a metric:

CCR (default): Credit Benchmark consensus only.
MyRating: adds your bank’s own rating and PD alongside consensus, for contributing banks.
SP: adds the S&P long-term foreign-currency rating alongside consensus.

Prompts

The server also provides two ready-made prompts. In clients that support MCP prompts, such as Claude, they appear in the prompt or attachment menu.

Prompt	Input	What it does
`cb_entity_brief`	A company name or LEI	Matches the entity, confirms the match, and summarises its consensus rating, recent trend, and rating changes.
`cb_portfolio_review`	Comma-separated company names, LEIs, or `CB_ID`s, plus an optional number of months (default 12)	Matches every name, reports any it could not match, then reviews the credit quality of the portfolio as a whole.

Interpreting results

PDs are decimal probabilities. 0.0034 means 0.34%, or 34 basis points.
Ratings default to the CB21 scale. Ask for CB7 or your bank’s master scale if you prefer. PDs are the same whichever scale is used.
Large results come back as CSV files. When a result is too large to show in the chat, the server returns the complete result as a downloadable CSV file.
Entity names link to the Web App. Select a name to open the entity’s page in the Credit Benchmark Web App.
Latest data by default. Results use the latest publication unless you ask for a specific date.

For field definitions, see the Data Dictionary. For how ratings and PDs are calculated, see the Methodology.

​Tools

​Prompts

​Interpreting results

Tools

Prompts

Interpreting results