Negative matching of JWT claims

I would like to allow JWT's that contain a claim that does NOT match a given value.

The problem I'm trying to solve is to disallow Dependabot to access certain secrets.

Example:

```
resource "vault_jwt_auth_backend" "github_jwt_auth" {
description = "JWT-based auth for GitHub Actions"
path = "jwt"
oidc_discovery_url = "https://token.actions.githubusercontent.com"
bound_issuer = "https://token.actions.githubusercontent.com"
}

resource "vault_jwt_auth_backend_role" "github_jwt_auth_role" {
backend = vault_jwt_auth_backend.github_jwt_auth.path
role_name = "actions-role"
token_policies = ["github_actions"]
bound_claims = {
"repository" : "KittyCAD/gha-jwt-test",
"runner_environment" : "github-hosted", # or "self-hosted"
"repository_visibility" : "github-hosted",
"actor": "dependabot[bot]"
}
user_claim = "actor"
role_type = "jwt"
}
```

The above is only giving dependabot access. However, I'd like to do the reverse.

No implementation files or tests are named. Start by locating the existing bound_claims matching logic and its tests, then determine how a negative claim condition should be configured and validated; done means the requested reverse Dependabot access behavior is covered by tests.

Written by the indexing model from the issue text.