Gigabrain

Gigabrain Memory Passport

An audit report for what your AI agents remember.

It scans the memory and instruction files your local agents already use, shows where each memory came from, flags risky or stale rows, and writes safe handoff briefs for the next agent.

In plain English: it is a local report plus export pack for AI memory. Think “security audit + travel document” for Codex, Claude Code, Cursor/Windsurf, OpenClaw, and manual cloud-memory exports.
Run a Passport See what it does

What is a Memory Passport?

A Memory Passport is the artifact you create before moving agent context between tools, teams, or repos. It answers three questions: what do the agents remember, can we trust it, and what is safe to hand off?

It scans local memory surfaces

Gigabrain reads visible local sources such as Codex memories, Claude Code project memory, Cursor/Windsurf rules, OpenClaw memory, and explicit manual exports from cloud assistants.

It produces a Passport folder

  • a Markdown/HTML/JSON audit report
  • a readiness score and blocker list
  • source inventory and provenance links
  • safe AGENTS.md and CLAUDE.md handoff briefs

It catches memory risk

The report flags duplicate memories, stale facts, contradictions, missing provenance, and secret-like rows before those memories are reused by another agent.

It does not scrape closed clouds

ChatGPT, Claude.ai, Gemini, and Copilot flows are manual import/export only. Secret-risk rows are never placed into generated handoff briefs.

The output is a folder you can inspect, share, or gate in CI.

The Passport report is static and local. You can open it in a browser, attach it to a review, or use the generated briefs to carry only approved context into the next agent.

Gigabrain Memory Passport report showing readiness, source inventory, and audit sections
Source inventory

See Codex, Claude Code, Cursor/Windsurf, OpenClaw, and manual cloud-export memory inputs in one local report.

Audit before handoff

Find duplicates, contradictions, stale facts, provenance gaps, and secret-risk rows before context moves again.

Safe briefs

Generate AGENTS.md, CLAUDE.md, ChatGPT, Claude.ai, Gemini, and Copilot briefs with secret-risk rows omitted entirely.

Run it locally.

The core is a CLI and MCP-backed control plane. Start with a static Passport, then graduate to policy checks, CI, and team review workflows.

npm install @legendaryvibecoder/gigabrain
npx gigabrainctl sync-hosts \
  --config ~/.gigabrain/config.json \
  --host codex,claude_code,cursor,windsurf
npx gigabrainctl passport \
  --config ~/.gigabrain/config.json \
  --output-dir ./gigabrain-passport

Built for teams using more than one AI coding agent.

The pilot turns invisible agent memory drift into an audit artifact your engineering and security teams can actually discuss.

Open core Local Passport

CLI, MCP tools, host sync, static report, safe handoff briefs, and local audit checks.

Design partner Team workflow

First audit, policy preset, remediation plan, and CI-ready Passport gate for one active repo.

Enterprise path Memory trust plane

Signed Passport attestations, admin review queues, policy packs, SSO, and fleet sync status.