Skip to main content
GRANWEL
GRANWEL ENTERPRISE SYSTEMS SYSTEM REV 2026

Infrastructure for the autonomous enterprise.

Granwel develops enterprise software for securing, governing, and understanding organisations operating with artificial intelligence.

Autonomous Governance
Execution sandboxing & policy gates
Institutional Memory
Decision provenance & knowledge graph
Cryptographic Audit
Tamper-evident SHA-256 hash chains
Static Surface
Zero-compromise security-by-design
Core Product Systems

Software architectures built for machine autonomy.

Granwel develops focused enterprise platforms that provide sovereign control, deterministic safety, and deep contextual memory to organizations executing with AI agents.

AEGIS by Granwel Developer Preview
Govern and secure autonomous engineering.

AEGIS is the execution governance and security kernel for autonomous AI software engineering. It intercepts agent actions, enforces deterministic policy boundaries, verifies architecture changes, and maintains an immutable cryptographic audit record.

  • Agent Governance: Strict identity and tool execution bounds.
  • Execution Control: Fail-closed mutating action gates.
  • Policy Enforcement: Machine-verifiable architectural ADR rules.
  • Auditability: Cryptographic hash-chained JSONL proofs.
AEGIS Core Architecture & Execution Flow
SPEC: KERNEL-ADAPTER-V1
1. AGENT INTERFACE 2. ADAPTER LAYER 3. AEGIS KERNEL 4. GOVERNED REPO Autonomous Agent Claude / Codex / Custom Developer Intent Directives & Scopes UNGOVERNED RISK: • Unchecked file mutates • Scope escape • Silent hallucination OpenCode Adapter Runtime Native Tool Intercept IDE / CLI Plugs Codex & Claude Hook Session Telemetry Policy Engine (aegis.yaml) ADR Scope • High-Risk Gate • Fails Closed Consensus & Evidence Core Multi-Agent Review • Risk Matrix Deterministic Pre-flight Tests Cryptographic Audit Chain Append-only JSONL SHA-256 Merkle Provenance Verified Mutate Bounded File Edits Certified Commit Proof-of-Verification Audit Proof Log .aegis/audit.jsonl Tamper-Evident
ATTICUS by Granwel Developer Preview
Organisational intelligence & institutional memory.

Atticus provides enterprise teams with a persistent, intelligent memory of everything they know. By constructing entity relationship graphs and decision provenance trees across repositories, ADRs, and communications, Atticus eliminates institutional knowledge loss.

  • Organisational Memory: Multi-source institutional ingestion.
  • Knowledge Retrieval: Sub-second hybrid relational & vector search.
  • Relationship Intelligence: Cross-domain team dependency mapping.
  • Decision Provenance: Unbroken historical rationale tracking.
ATTICUS Organisational Intelligence Flow
SPEC: KNOWLEDGE-PROVENANCE-V1
1. INGESTION PIPELINE 2. KNOWLEDGE CORE 3. PROVENANCE ENGINE 4. VERIFIED CONTEXT Enterprise Repositories Commits, PRs, Discussions Architectural ADRs Decisions & Technical Scopes Comms & Incidents Postmortems & Rationale CHALLENGE SOLVED: • Fragmented institutional loss • Silent context drift Rust / Axum Backend High-throughput ingestion Strict schema validation Relational & Vector Hybrid retrieval matrix Deterministic index Entity Link Graph Author ↔ Code ↔ Decision Topological mapping Decision Provenance Full lineage of architectural choices Why was this changed? By whom? Relationship Intelligence Team domain boundaries Cross-system dependency risk Institutional Memory Bus Cryptographically referenced Immutable context audit Semantic Query API Sub-second retrieval Exact citation links Verified Feeds Agent context injection Zero ungrounded noise Enterprise RBAC Identity partition Tenant isolation
Platform Principles

Independent architectures. Unified enterprise primitives.

Granwel products are built to operate independently while interoperating through a common foundation of cryptographic proofs, deterministic policy controls, and unified enterprise primitives.

Granwel Enterprise Platform Architecture
SPEC: GRANWEL-UNIFIED-PRIMITIVES-V1
AEGIS BY GRANWEL Autonomous Engineering Governance Agent Sandboxing • Policy Gates • Hash-Chained Audit ATTICUS BY GRANWEL Organisational Intelligence Platform Institutional Memory • Decision Provenance • Knowledge Graph GRANWEL COMMON ENTERPRISE PRIMITIVES Cryptographic Provenance Execution Policy Engine Institutional Memory Bus Fail-Closed Governance

Cryptographic Provenance

Every decision, mutating action, and context query is cryptographically referenced with append-only hash chains for auditability.

Fail-Closed Governance

Unverified agent instructions, missing authorizations, or unapproved architectural mutations are rejected by default.

Zero Untrusted Dependency

Granwel systems are engineered with minimal attack surfaces, isolated memory spaces, and sovereign deployment models.

Security Philosophy

Security by design. Defensible engineering.

We reject the industry practice of exaggerated marketing claims. Security cannot be reduced to buzzwords. At Granwel, security is an architectural discipline governed by formal threat modeling, verifiable supply chains, and strict isolation boundaries.

Static Production Attack Surface ZERO RUNTIME CMS/DB
Content Security Policy STRICT SELF-HOSTED
Third-Party Trackers & Scripts NONE PERMITTED
RESPONSIBLE SECURITY COMMITMENT

Granwel maintains an open and transparent security disclosure process under RFC 9116. If you discover a potential vulnerability across any Granwel repository or infrastructure, our security engineering team processes reports directly.

# Direct Security Contact
Institutional Access

Engage with Granwel Engineering

Granwel collaborates with enterprise engineering leadership, autonomous systems researchers, and security teams seeking deterministic control over agentic software workflows.