Cognitive architecture for Claude Code – triggers, memory, docs
Trigger-based cognitive architecture for Claude Code loses context anyway without API-level state persistence.

Agent-authored docs reviewed in PRs beat stale Confluence pages.
Engineering teams struggling with outdated architecture documentation
Structurizr · Swimm · Mintlify
Trigger-based cognitive architecture for Claude Code loses context anyway without API-level state persistence.
AGENTS.md with retrieval discipline — Cursor rules and Continue.dev already provide similar context.
Blast radius detection before AI edits code, competing with Cursor's codebase awareness.
Surfaces architectural decisions from git history — Cursor and Continue don't do this.
Turns subjective docs quality into objective pass/fail metrics using dumb agent harnesses.
Drag-drop sitemaps with AI hints—but Figma, Slickplan, and Whimsical already own this category.