Agentic Bookmarks

Security Overview

Last updated: May 12, 2026

Security Overview

Effective: May 12, 2026 Last updated: May 12, 2026 Canonical URL: https://agenticbookmarks.com/legal/security

At a glance

Agentic Bookmarks is a VS Code extension and bundled MCP server designed so that the easiest answer to most enterprise security questions is: the data isn't there to begin with.

  • No telemetry. No analytics, no behavioral logging, no usage events.
  • No source code, prompts, or repository data leaves your machine.
  • AI features run locally using your own AI provider. We never see prompts or completions.
  • Repository visibility is checked from your machine, not via our servers.
  • Open and inspectable. The public-repository portions of the Software, including the entire network layer, are published as source-available code at https://github.com/super-mega-lab/agentic-bookmarks under PolyForm Shield 1.0.0 so you can verify our network behavior, not just trust it.

If you're filling out a security questionnaire about this product, this page is intended to be the answer to most of it. If anything is unclear or insufficient, email contact@supermegalab.com.

1. Architecture and data flow

1.1 What runs where

ComponentLocationPurpose
VS Code extensionUser's deviceUI, repository operations, feature gating
MCP serverUser's deviceLocal protocol bridge for AI agents
Repository hosts (GitHub, GitLab, etc.)Third-party infrastructureRepository operations and visibility checks, contacted directly from the user's device
User's AI providerUser's choiceAI/MCP interactions, contacted directly from the user's device

1.2 What flows where

                        ┌────────────────────────┐
   User's device        │                        │
 ┌──────────────────┐   │   Repository hosts     │
 │                  │◄──┤  (GitHub, GitLab, …)   │
 │  VS Code         │   │                        │
 │  Extension       │   └────────────────────────┘
 │                  │
 │  MCP Server  ────┼──►┌────────────────────────┐
 │                  │   │   User's AI provider   │
 │                  │   └────────────────────────┘
 │                  │
 │  Network layer   │
 │  (open source)   │
 │                  │
 └──────────────────┘

 No arrows to Provider infrastructure: the Software does not contact
 Super Mega Lab during the Beta Period.

1.3 Data flow to Provider

None. During the Beta Period the Software does not contact Super Mega Lab's infrastructure at all. We do not receive source code, prompts, AI interactions, repository names, branch names, file paths, commit metadata, telemetry events, error reports, repository visibility status, IP addresses, or any other content from the user's development environment.

2. Verify, don't trust

A useful property of the architecture: you can verify these claims yourself.

The portions of the extension that contain any network-capable code are published as source under PolyForm Shield 1.0.0 at https://github.com/super-mega-lab/agentic-bookmarks. You can audit those components to confirm:

  • That the Software does not contact Super Mega Lab during the Beta Period — the outbound host allowlist excludes any Provider service.
  • That repository content, prompts, telemetry, or other development-environment data are never sent to Provider.
  • That the repository visibility check runs on the user's device by querying the applicable repository host directly.

The proprietary core of the Software (the implementation of paid Pro features and the shared logic both the extension and MCP server depend on) is bundled with the Marketplace release as a compiled artifact. The proprietary core does not initiate network requests; all outbound network communication is the exclusive responsibility of the open network layer in the public repository.

3. Encryption and data protection

The Software does not transmit data to Provider infrastructure during the Beta Period. Where the Software contacts third-party services on your direction (repository hosts, your AI provider), it uses the encrypted endpoints those services provide (HTTPS / TLS).

4. Authentication and access control

During the Beta Period, all Pro features are available to all users without a subscription, and the Software does not authenticate against any Provider service. No account or sign-in is required.

5. Network design

The Software enforces an explicit allowlist of outbound hosts. The allowlist is implemented in the public network layer at https://github.com/super-mega-lab/agentic-bookmarks. During the Beta Period the allowlist excludes Provider infrastructure entirely. Any change to the allowlist requires a code change visible in the public repository.

6. Logging and monitoring

Provider operates no service that the Software contacts during the Beta Period, and therefore Provider does not log any user activity, IP address, request, or development-environment data.

7. Vulnerability management

7.1 Dependencies

The published source of the network layer at https://github.com/super-mega-lab/agentic-bookmarks enables independent inspection of the Software's third-party dependencies. Dependencies are reviewed for known vulnerabilities prior to each Marketplace release. Critical and high-severity issues block release until remediated.

7.2 Code review

Production code is reviewed prior to release. Releases are produced from the public repository's release branch.

7.3 Secret scanning

The public repository enables automated secret scanning at the hosting provider level (GitHub secret scanning).

8. Personnel

  • All personnel with access to systems handling personal data sign written confidentiality obligations.
  • Security and data-protection training is provided on hire and annually thereafter.
  • Background checks are performed where lawful and appropriate to the role.

9. Sub-processors

Provider engages no sub-processors to process user data during the Beta Period, because Provider receives no user data.

10. Incident response

We maintain a documented incident response procedure that covers detection, triage, containment, eradication, recovery, and post-incident review.

In the event of a security incident affecting personal data, we will notify affected customers without undue delay, and in any event within seventy-two (72) hours of becoming aware, consistent with our DPA and applicable law. Notification will include, to the extent known: nature of the incident, approximate scope, measures taken or proposed, and a contact for further information.

11. Responsible disclosure

If you believe you have found a security issue in the Product or our services:

  • Email contact@supermegalab.com.
  • Please include enough information to reproduce the issue.
  • We will acknowledge receipt within five (5) business days and keep you informed of our progress.
  • We commit not to pursue legal action against good-faith security researchers who follow this disclosure process and avoid privacy violations, service disruption, and data destruction.

12. Data subject rights

You can exercise the data protection rights described in the Privacy Policy (access, rectification, erasure, portability, restriction, objection, complaint) by emailing contact@supermegalab.com. Given the limited data we hold, we are usually able to respond within 30 days at no charge.

13. Compliance

We have designed our security program against the principles in ISO/IEC 27001 and SOC 2 Trust Services Criteria. We are not yet certified or attested under either framework.

We comply with applicable data protection laws including the GDPR, UK GDPR, FADP, and the CCPA/CPRA, as described in the Privacy Policy and DPA.

14. International data transfers

Where applicable, we rely on the EU Standard Contractual Clauses, the UK International Data Transfer Addendum, and equivalent Swiss safeguards for cross-border transfers of personal data, as described in the Privacy Policy.

15. Things we deliberately don't do

These are commitments, not omissions:

We don'tWhy
Telemetry, analytics, or behavioral event loggingPrivacy-by-design
Receive or process source code, prompts, or repository contentArchitectural
Train AI/ML models on customer data of any kindPolicy + architectural — no data flow exists
Sell, share, rent, or trade personal informationPolicy
Use third-party advertising or marketing trackersPolicy
Run AI features through our serversLocal-first architecture

If we ever change one of these, it is a material change under our Privacy Policy and Provider-Specific Terms requiring at least 30 days' notice and, where applicable, opt-in consent.

16. Documents and links

17. Contact