Vercel API handlers more often return JSON than HTML, so res.send is
not the only response body sink that matters. Mirror Express's
ResponseJsonCall by also matching res.json(...) and res.jsonp(...) on
the response (direct and chained), and exercise the new behavior in
the library-test fixture.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Fixes US spelling (recognised -> recognized) across docs, QLDoc,
change note, and test fixture comments. Clarifies the handler QLDoc
to note sync/async support. Renames the supported-frameworks entry
from "vercel" to "Vercel (@vercel/node)" to avoid implying broader
platform coverage.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Extends the Vercel serverless handler detection to also match the
deprecated Zeit-era @now/node package with NowRequest/NowResponse
types. Per-review feedback from asgerf, these aliases still appear
in real-world code.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This adds a framework model for Vercel serverless functions so that
CodeQL's existing JavaScript security queries can detect vulnerabilities
in handlers of the form
export default function handler(req: VercelRequest, res: VercelResponse) { ... }
Handlers are identified as the default export of a module whose first
two parameters are typed as `VercelRequest`/`VercelResponse` from
`@vercel/node`. The default-export constraint excludes private helpers
that share the same signature. Type-based detection follows the same
pattern already used by `NextReqResHandler` in `Next.qll`.
The framework model covers:
- Route handler recognition (default-exported typed handlers only)
- Request input sources: `query`, `body`, `cookies`, and `url`
(the last inherited from Node's `IncomingMessage`)
- Named header accesses like `req.headers.host` and `req.headers.referer`,
modelled as `Http::RequestHeaderAccess` so header-specific queries fire
- Response sinks: `res.send`, `res.status(...).send`, `res.redirect`
- Header definitions via `res.setHeader`
Includes a library test exercising each model predicate (including a
negative case for private helpers) and query consistency fixtures
demonstrating end-to-end detection for js/reflected-xss,
js/request-forgery, js/sql-injection, and js/command-line-injection.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Fix misplaced semicolons in test files (was inside comment, moved before it)
- Update QLdoc comments to reference new browser source kind names
- Update docs to list browser source kinds and fix outdated 'only remote' note
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
A bug made it into the release which causes compilation errors when
forceLocal is used on a predicate with a result column.
This commit works around the issue by converting the result column
to a positional parameter, for the predicates that we use forceLocal on.
It should be safe to revert this commit once the compiler fix has made
it into a stable release.
Some abstract classes defines fields without binding them, leaving it up to the subclasses to bind them. When combined with overlay[local?], the charpred for such an abstract class can become local, while the subclasses are global. The means the charpred needs to be materialized, even though it doesn't bind the fields, leading to a cartesian product.
