Why I Built the "Infrastructure Layer" Under Every AI Coding Agents

A developer built ASIL (Engineering Intelligence Infrastructure), a persistent knowledge graph that connects code, dependencies, runtime behavior, and incident history into a queryable system accessible by any AI agent through MCP. The infrastructure addresses the limitation that AI coding agents lack persistent understanding of a system, forcing them to rediscover the same engineering knowledge repeatedly. ASIL stores verified engineering conclusions in persistent memory, enabling agents to reuse prior reasoning on cache hits at near-zero cost rather than re-running full LLM pipelines.

AI coding agents are getting very good at editing files, running tests, and opening PRs. After heavily using tools like Cursor, Claude Code, and GitHub Copilot, I noticed they all share the same core limitation: They have no persistent understanding of your system. Ask the same question next week and they: So instead of building another coding agent, I built the layer underneath them. ASIL Engineering Intelligence Infrastructure is a persistent, temporal, causal knowledge graph for software systems. It connects: into one queryable system that any AI agent can access through MCP. The goal is simple: Stop making AI agents rediscover the same engineering knowledge over and over again. Most coding agents understand: ASIL understands: Instead of: “GPT thinks this caused the outage” ASIL derives causal chains from observable system state: Every conclusion includes: No black-box “AI intuition.” uv run asil ask "How does auth work in this repo?" ASIL combines: to return: uv run asil replay INC-2026-04-12 ASIL reconstructs: as a dependency-aware replay graph. Think: Time-travel debugging for distributed systems. uv run asil drift report ASIL learns expected dependency boundaries and flags: before the PR merges. ASIL exposes 13 MCP tools usable from: The agents become clients of the intelligence layer. ASIL stores every verified engineering conclusion in persistent memory. When someone asks a semantically similar question later, ASIL can reuse the prior verified reasoning instead of re-running the full LLM pipeline. On cache hits, the cost drops close to: just the embedding lookup Repeated engineering queries become dramatically cheaper over time — especially across teams. ASIL does not let the LLM invent causality. That rule shapes the entire architecture. Causal links come from deterministic signals: The LLM consumes evidence. It does not fabricate it. That distinction matters once AI systems start participating in production engineering workflows. Everything runs locally: No central server. No telemetry. Your graph stays yours. The only optional network dependency is the reasoning LLM. Most AI tooling is racing toward: “make the agent better at editing code” I think the more important problem is: “give agents persistent engineering intelligence” That means: That’s the layer ASIL is trying to build. Built solo over 6 months with: Python, FastAPI, Neo4j, Qdrant, Postgres, Tree-sitter, Next.js, Tailwind, ReactFlow, and MCP tooling.