You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit 1c9b296
Browse filesBrowse the repository at this point in the historyBrowse files
Python: Use the ai threat model for MCP handler input
Handler parameters are produced by a language model, so they are sources of the opt-in ai threat model added in github#22810. HTTP headers and bearer tokens come from the HTTP client and stay sources of the remote threat model.
* Added modeling of the `mcp` (the official Model Context Protocol SDK) and `fastmcp` PyPI packages. The parameters of MCP tool, resource and prompt handlers, the HTTP headers that `fastmcp` handlers read through `get_http_headers()` or `get_http_request()`, and the bearer token that token verifiers (`verify_token`) and `get_access_token()` receive, are now sources of the `remote` threat model.
4
+
* Added modeling of the `mcp` (the official Model Context Protocol SDK) and `fastmcp` PyPI packages. The parameters of MCP tool, resource and prompt handlers are now sources of the opt-in `ai` threat model, as their values are produced by a language model. The HTTP headers that `fastmcp` handlers read through `get_http_headers()` or `get_http_request()`, and the bearer token that token verifiers (`verify_token`) and `get_access_token()` receive, are sources of the `remote` threat model.
0 commit comments