Skip to main content
AEGIBIT
Security

Security is the architecture. Not the marketing.

Every claim on this page is defensible against a CISO's first question: "where does that number come from?"If we can't point at source code, a configuration file, or a public artifact, the claim isn't on this page.

The posture

Six things AEGIBIT actually does, each anchored to a file you can read or a public artifact you can audit.

01

Hardened response headers on every request

Explicit Content Security Policy with no unsafe-eval. HSTS with includeSubDomains. Permissions-Policy denies camera, microphone, geolocation, payment, sensors. COOP same-origin-allow-popups. X-Content-Type-Options nosniff. Pinned in next.config.ts and verified on every deploy.

↳ proof: next.config.ts headers() function

02

Per-IP rate limiting that survives serverless cold starts

Upstash Redis-backed limiter (visitor 10/min, event 30/min, lead 5/min, chat 20/min). In-memory fallback when Upstash unreachable. Fail-open posture so a vendor outage doesn't lock out legitimate visitors. Lambda spawn doesn't reset the counter.

↳ proof: src/lib/rate-limiter.ts

03

httpOnly cookie sessions, scrypt-hashed admin credentials

Iron-session encrypted cookies (SameSite=Lax, Secure in production, 8-hour TTL). Admin password stored as scrypt:salt:derived format with N=2^14 and 64 MB maxmem, explicitly bounded so a misconfigured runtime can't OOM. No bearer tokens in client bundles, ever.

↳ proof: src/lib/auth.ts + src/lib/session.ts

04

Append-only audit trail on every automation action

Every agent run, every deploy notify, every lead-pipeline action writes a row to a Postgres table with start time, finish time, status, full payload, and the agent identity. Rows are insert-only; no UPDATE or DELETE path in the application code.

↳ proof: supabase/schema.sql `agent_actions` table

05

PayMint voucher capture, auditor-grade from the field

Every voucher captured at the branch is anchored to a photograph, a GPS coordinate, and a server-issued timestamp at the moment of submission. The visibility window for headquarters is same-day instead of the 5-9 days typical of paper-based reconciliation.

↳ proof: PayMint product feature, see /products/paymint

06

MCP Shield, we open-sourced our own security tooling

After Anthropic's April 2026 MCP-RCE disclosure, we shipped MCP Shield: a free, MIT-licensed scanner + runtime for Model Context Protocol servers. Detects tool poisoning, hidden-Unicode steganography, prompt injection markers, secret exposure, unsafe stdio launches. Public code, public issues, public reports.

↳ proof: github.com/AegibitSecurity/mcp-shield + /products/mcp-shield

Compliance status, today

We will not display a certification badge we have not earned.

DPDP Act 2023 (India)

Aligning

AEGIBIT is an Indian company processing Indian PII. The Digital Personal Data Protection Act 2023 applies by default. We process lead-form data with consent, store only what's necessary, and treat the user's email as the primary identity reference.

SOC 2 Type II

In progress

Internal controls + audit-trail discipline mapped to the Trust Services Criteria. Formal Type II audit window has not yet opened. We will not display a SOC 2 badge until the auditor issues the report.

GDPR, EU visitor data

Aware

EU-resident visitors who submit forms can request data export or deletion via contact@aegibit.com. We do not currently market into the EU; if AEGIBIT begins selling there, we will publish a formal DPA.

What we do NOT claim

  • · ISO 27001, not certified
  • · HIPAA, not applicable (we are not a US healthcare entity or covered associate)
  • · RBI Cybersecurity Framework, we are not a regulated bank or NBFC
  • · SEBI CSCRF, not applicable
  • · FedRAMP, IRAP, C5, not pursued

If your buying process requires a specific certification that's not on the "today" list above, the AEGIBIT team can share our Trust Center roadmap, reach out at contact@aegibit.com.

Responsible disclosure

Found a vulnerability? Tell us, we'll fix it and credit you.

Email contact@aegibit.com with subject "Security Disclosure." Include a clear reproduction path, the affected URL or component, and any payload required to trigger the issue. Encrypted channels available on request.

We commit to acknowledging the report within 48 hours, providing a remediation timeline within 7 days, and coordinating a public disclosure window with you once a fix has shipped. We will not pursue legal action against anyone reporting in good faith.

Out of scope: denial-of-service testing against production traffic, social engineering of AEGIBIT personnel, and any test that compromises the data or experience of other AEGIBIT visitors or customers.

Want the security walkthrough live?

Book a 12-minute PayMint demo and we'll cover the security architecture alongside the product workflow. Configuration files open, no slideware.