PAPERS

Technical writing on Forge.

Architecture deep-dives, comparisons against existing tooling, security posture, and design rationale. Engineering substance, not marketing material.

FEATURED
WHITEPAPER · TECHNICAL · 6+ PAGESJune 2026

Forge: Architecture & Engineering

How Forge generates personalized AI agents through a Phase 1.5 Context Interview, the 32-capability catalog, the 96-tool runtime surface, self-modification mechanics, security model, and comparisons against LangChain, raw LLM APIs, and building from scratch.

10 sections·PR citations·PDF-friendly
Read the whitepaper →
COMPANION READS
FOR ENGINEERS

Forge vs LangChain vs DIY

Comparison tables. Line counts. PR ledger. Why Forge is a generated agent vs LangChain abstractions you assemble. For people who'd rather not spend 2-3 months on glue code.

For developers →
FOR OPERATORS

What you can actually build

Sales chief of staff. Investment thesis agent. Personal CTO. Operations brain. Six example agents with the concrete things they actually do for you.

For operators →
REFERENCE
SECURITY POSTURE

Honest trust model, verified controls, known limitations.

Adversarial-audited sandbox boundaries. What we verify with tests, what's intentional design, what's a known limitation. References PR #113 audit findings. No marketing language.

Read the security posture →
PLANNED
  • Q3 2026Capability deep-dive: persistent_memory & A-MEM (graph notes, fastembed ONNX, retrieval decay)
  • Q3 2026Self-modification: how code_modification + restart_self compose with the workspace freshness check
  • Q4 2026The 32-capability catalog: design rationale, what we cut, what we added, what we'd cut next
  • Q4 2026Phase 1.5 Context Interview: a 5-wave conversational design protocol for capturing user voice and constraints

Want to skip the reading and build the thing?

Design an agent