Autocomplete & Security

Autocomplete respects your Bonsai instance's security policy. Properties that are blocked at evaluation time are also blocked from appearing in suggestions.

allowedProperties

When your instance uses allowedProperties, only those properties appear in completions. Everything else is excluded.

const expr = bonsai({ allowedProperties: ['name', 'age'] })

const ac = createAutocomplete(expr, {
  context: { user: { name: 'Alice', age: 25, role: 'admin' } },
})

ac.complete('user.', 5)
// [{label: 'name', ...}, {label: 'age', ...}]
// 'role' is not suggested because it's not in the allowlist

deniedProperties

When your instance uses deniedProperties, those properties are excluded from suggestions. Everything else is allowed.

const expr = bonsai({ deniedProperties: ['password', 'secret'] })

const ac = createAutocomplete(expr, {
  context: { user: { name: 'Alice', password: 'hunter2' } },
})

ac.complete('user.', 5)
// [{label: 'name', ...}]
// 'password' is never suggested

Lambda and pipe contexts

Security filtering applies everywhere: top-level property access, lambda element properties, pipe transform suggestions, and method completions. There is no context where a blocked property can leak into suggestions.

const expr = bonsai({ allowedProperties: ['name'] }).use(strings)

const ac = createAutocomplete(expr, {
  context: {
    users: [{ name: 'Alice', age: 25, role: 'admin' }],
  },
})

ac.complete('users.filter(.', 15)
// [{label: 'name', ...}]
// 'age' and 'role' are excluded from lambda suggestions too

Security policy is fixed at construction time

The security policy is read once when createAutocomplete() is called and cached for the lifetime of the autocomplete instance. If you need different policies for different users or roles, create separate autocomplete instances.