Signaturesi
Trust Security

Security without empty promises.

Signaturesi is being designed around controlled access, account isolation, responsible data handling and clear disclosure of current security limitations.

Updated July 24, 2026
Version 1.0 Draft
Scope NEO, Bean and Bean ID
Current security position Signaturesi is in active development.

Some safeguards described on this page are product requirements rather than verified production claims. Final wording must match the deployed architecture, completed testing and independent security review.

Development
01

Security should be designed into the product.

Signaturesi is building connected services for intelligence, communication and unified identity. These services may process account information, conversations, messages, files and personal settings.

Security therefore cannot be treated as a visual feature or marketing label. It requires secure authentication, strict access rules, careful data handling, reliable monitoring and a documented response process.

Core position We will describe a protection as implemented only after it has been built, tested and verified.
02

Six principles guide Signaturesi security.

Minimum access

Systems and people should receive only the access required for a defined purpose.

Secure defaults

New accounts and features should begin with conservative access and privacy settings.

User isolation

One user must not be able to access another user’s private account or application data.

Continuous visibility

Important authentication, access and system events should be logged and monitored.

Limited exposure

External providers should receive only the information necessary for a requested function.

Recovery readiness

Security failures should trigger containment, investigation, recovery and user communication.

03

One identity requires strong account protection.

Bean ID connects access across Signaturesi products. A compromised Bean ID could affect NEO, Bean and related account data, making authentication security a critical requirement.

Bean ID Unified identity
Required controls
Password protection

Passwords should be stored using a trusted password-hashing system, never as readable text.

Session management

Sessions should expire, rotate appropriately and support revocation after suspicious activity.

Rate limiting

Login, signup and recovery endpoints should resist automated guessing and abusive requests.

Account recovery

Recovery should verify account ownership without exposing secrets or enabling easy takeover.

Device visibility

Users should be able to review and remove active authenticated devices or sessions.

Security notifications

Important login, recovery and account-change events should generate clear alerts.

Multi-factor authentication

Multi-factor authentication is a recommended production requirement. It should not be described as available until the complete enrollment, recovery and backup-code flows have been implemented.

04

AI systems need isolation between users, memories and tools.

NEO may process prompts, generated responses, uploaded files, project information and personal memory. Security controls must prevent unauthorized users, tools and services from accessing unrelated information.

NEO AI workspace
Security model
Layer 01
Account boundary

Every conversation, file and memory item should be associated with the correct user.

Layer 02
Retrieval boundary

Memory retrieval should search only information the current account is authorized to access.

Layer 03
Tool boundary

Connected tools should receive only the minimum information needed for the current task.

Layer 04
Action confirmation

Sensitive external actions should require explicit user confirmation.

Prompt injection and untrusted content

Uploaded files, websites and retrieved content may contain instructions intended to manipulate an AI system. NEO should treat external content as untrusted data and should not automatically allow it to override system permissions or user intent.

Memory integrity

Durable memory should not be created silently from every conversation. Users should be able to inspect, correct and delete saved memory, and sensitive data should be excluded or restricted by default.

05

Communication security covers more than message content.

Bean may process message content, attachments, presence information, delivery records and call connection data. Each part requires its own access, retention and security controls.

Bean Communication
Security areas
Messages

Delivery and storage access should be limited to authorized conversation participants.

Media

Files should use private storage paths, access checks and controlled download links.

Presence

Online and last-active information should follow user-selected privacy settings.

Calls

Signaling, permissions and connection metadata should be protected from unauthorized access.

Abuse prevention

Bean should include controls for blocking, reporting, spam reduction, suspicious account detection and rate limiting without exposing private conversations more broadly than necessary.

06

Encryption claims must match the real implementation.

Signaturesi intends to protect data during supported network transmission and to use appropriate storage protections through its infrastructure providers.

Intended

Encryption in transit

Supported production traffic should use secure HTTPS or equivalent encrypted transport.

Provider dependent

Encryption at rest

Database, storage and backup protection depends on the final infrastructure configuration.

Not verified

End-to-end encryption

Bean must not claim end-to-end encryption until message, media, backup and call protection has been fully implemented and independently reviewed.

Approved wording during development Bean is being designed with privacy, account isolation and controlled access as core product requirements.
07

Configuration matters as much as provider selection.

Signaturesi may rely on external hosting, database, network, communication and AI providers. Using a reputable provider does not automatically make an application secure.

Database Row-level authorization

Database policies should prevent cross-account reads, writes and unauthorized administrative access.

API Server-side validation

Sensitive authorization must not rely only on checks performed in browser JavaScript.

Storage Private file access

User files should not be exposed through permanent public URLs unless explicitly intended.

Secrets Protected credentials

API keys, private tokens and service credentials must not be embedded in public frontend files.

Network Traffic protection

Rate limiting, abuse detection and secure transport should protect public endpoints.

Backups Tested recovery

Backups should follow defined retention, access and restoration procedures.

Never expose private service keys

Secret API keys, database service-role keys and administrative credentials must remain in secure server-side environments.

08

Every sensitive request must be authorized.

Authentication confirms who is making a request. Authorization determines what that person is allowed to access. Signaturesi requires both.

Step 01
Authenticate

Verify the account and active session.

Step 02
Validate

Confirm request structure and permitted values.

Step 03
Authorize

Confirm the account owns or may access the resource.

Step 04
Record

Log important events for security investigation.

Hidden buttons, inaccessible interface elements or private-looking URLs are not security controls. Authorization must be enforced by the backend.

09

Collect less, retain less and expose less.

Reducing unnecessary data can reduce security risk. Signaturesi should limit collection and retention to what is required for product functionality, security and legal obligations.

Collection Gather only required data

Do not request device permissions or personal information without a clear feature requirement.

Separation Keep account data isolated

User-owned information should remain separated through database and storage access rules.

Retention Define deletion periods

Temporary files, logs and deleted account data should follow documented retention schedules.

Exposure Limit provider access

External providers should receive only the data needed for a requested operation.

10

Security events need useful visibility.

Signaturesi should record security-relevant events without placing private message or conversation content unnecessarily into logs.

Event Why it matters Recommended response
Repeated failed login Possible credential guessing Rate limit and review
New recovery request Possible account takeover attempt Notify account owner
Unusual data access Possible authorization failure Block and investigate
Elevated error rate Possible outage or attack Trigger operational alert
Administrative action High-impact internal access Record actor and reason

Logs should themselves have restricted access and defined retention periods.

11

Security incidents require a defined process.

A security incident may include unauthorized access, leaked credentials, data exposure, malware, service abuse or a critical provider failure.

01
Detect

Identify suspicious activity through reports, alerts or system monitoring.

02
Contain

Revoke access, disable compromised credentials or isolate affected components.

03
Investigate

Determine affected systems, data, accounts and the likely cause.

04
Recover

Restore safe operation and verify that the vulnerability has been addressed.

05
Communicate

Notify affected users or authorities when appropriate or legally required.

06
Improve

Document the incident and prevent recurrence through technical and process changes.

View system status
12

Responsible security research is welcome.

Security researchers who believe they have discovered a vulnerability should report it privately and allow reasonable time for investigation and remediation.

Include
  • A clear description of the vulnerability.
  • The affected page, endpoint or product.
  • Reproduction steps or a safe proof of concept.
  • The possible security impact.
  • A contact method for follow-up.
Do not
  • Access or download other users’ private data.
  • Disrupt production services or availability.
  • Use social engineering against users or staff.
  • Deploy malware or destructive payloads.
  • Publicly disclose an unresolved vulnerability.

Report a vulnerability privately.

Replace this address with a verified and actively monitored security inbox before launch.

security@signaturesi.com

Signaturesi should publish a complete responsible disclosure policy before inviting external testing or offering a bug-bounty program.

13

Security work remains incomplete.

Until development and independent review are complete, Signaturesi should clearly identify unresolved security areas rather than hiding them behind broad claims.

Unverified
End-to-end encryption

Full protection across messages, media, backups and calls has not been publicly verified.

Pending
Independent penetration test

No completed external security assessment is claimed on this page.

Pending
Multi-factor authentication

Availability depends on the final production account implementation.

Pending
Formal security certifications

Signaturesi does not currently claim formal security compliance certification.

Required
Incident response testing

Recovery and notification procedures should be tested before production launch.

14

The next security work should be measurable.

Phase 01
Authentication hardening

Secure login, session expiry, recovery, rate limiting and device management.

Phase 02
Database isolation review

Test row-level policies, storage permissions and backend authorization.

Phase 03
Messaging security review

Review message delivery, media access, call signaling and metadata handling.

Phase 04
External penetration testing

Engage qualified independent reviewers before making stronger public security claims.

Phase 05
Incident-response exercise

Test detection, containment, recovery and user communication procedures.

15

Account security also depends on user behavior.

Users should take reasonable steps to protect their Signaturesi accounts and devices.

  • Use a strong password that is not reused on another service.
  • Keep account recovery methods accurate and protected.
  • Do not share passwords, login codes or recovery links.
  • Sign out from devices you no longer control.
  • Review unexpected login or account-change notifications.
  • Keep browsers, operating systems and devices updated.
  • Report suspected account compromise immediately.

Signaturesi will never need a user’s password through an ordinary support message.

16

Contact the Signaturesi security team.

Use a private and verified channel for vulnerability reports, account-security concerns or suspected incidents.

Signaturesi Security Vulnerabilities and incidents
Draft contact
Response target To be confirmed

These email addresses must be created, secured and actively monitored before this page is published.