PACT — Programmatic Agent Constraint Toolkit
A framework for governing AI coding agents through infrastructure instead of instructions.
The common assumption is that Claude's problem is memory. It forgets facts, decisions, and other nuances between sessions, so the go-to "solution" is to make it remember more by applying bandaids like session logs or plans in endless markdown files, classic databases, spawning dozens of subagents to argue it out with little context, and attempting marathon sessions to prove you've avoided compaction; that last one is actually nightmare fuel, not an accomplishment. I'll take "context rot" for $500, Alex.
The uncomfortable truth is that Claude makes bad decisions even when it remembers everything. It will edit a file it hasn't read. Fix one layer and break three downstream. Guess at a package API instead of checking docs. Declare "done" while half the work is missing. Full context doesn't fix any of these. They're reasoning failures, not recall failures.
Memory plugins ask: "How do we help Claude remember?" PACT asks: "How do we help Claude think?"
Who Is PACT For?
Solo developers and small teams. PACT was built by a solo developer for solo developers, then designed to scale up. The smaller your team, the more you need it. There's no code reviewer to notice the stale cache, no teammate to check the dependency chain, no QA to catch the UI regression. Large teams have humans for that. You have PACT.
Small project? Good. That's when infrastructure matters most. PACT's subsystems activate on the patterns they detect, not on project size.
What Is PACT?
PACT is a modular governance framework for AI coding agents (Claude Code, Cursor, Copilot Workspace, etc.). Use all of it or just the parts that fill gaps in your existing setup — if you already have a memory layer, a task manager, or a workflow orchestrator, PACT detects that and only scaffolds what's missing.
PACT has nine features. Take what you need:
- Mechanical Enforcement — Shell hooks that block violations before they land
- Context Replacement — Architecture maps and lifecycle flows that replace memory
- Self-Evolving Reasoning — Questions the agent asks itself at decision points
- Structure/Behavior Separation — Static wiring maps vs dynamic lifecycle flows
- Multi-Agent Resilience — When Claude is down, switch to Gemini (or vice versa) with zero context loss
- Compound Intelligence — Research synthesis, cross-system knowledge directory, and capability baseline that make each session smarter than the last
- Observability & Feedback — Real-time dashboard that visualizes agent activity, captures user prompts, tracks tasks, and feeds user ratings back into future sessions
- Distributed Cognition — Auto-dispatched subagents for dependency tracing, knowledge research, and pre-commit review so the main session stays focused on the user's task
- Vector Memory — Semantic search across bugs, solutions, research, and task feedback using local embeddings (no API keys, no cloud). YAML stays authoritative; vector search finds the right file faster
Multi-Agent Support (Claude + Gemini)
PACT supports running multiple AI agents on the same project — Claude Code and Gemini CLI share the same hooks, rules, and task tracker. When one agent is degraded, switch to the other without losing context.
- Shared governance: One set of rules (CLAUDE.md), one set of hooks, one task tracker
- Model identity: Sessions and commits are auto-tagged with the agent name
- Seamless handoffs: PENDING_WORK.yaml tracks what each agent was doing
- Status monitoring:
session-status-check.shwarns you when Claude is degraded
Read the full Multi-Agent Setup Guide →