Files drift across sessions. Anchors change but consumers don't. Docs fall behind.
TRACE enforces structural integrity — automatically.
Every AI-assisted session operates without memory of what depends on what.
AI modifies a core interface. Six files depend on it. They silently break.
Session crashes or token limit hits. The AI has no idea where you left off.
Tests removed. Complexity creeps in. Docs fall behind. Nobody notices until production breaks.
Automated, gate-enforced checks at every session boundary.
Full enforcement from day one, or gradual adoption with zero disruption.
Scaffolds everything. Gates default to "block." Define anchors, run gate start, build.
Auto-detects anchors, consumers, tests, quality tools. Old code gets a baseline pass. Only new code enforced.
Validates state, integrity, debt.
Generates AI context.
Code with checkpoints.
Auto-observation captures changes.
5-pillar check, quality tools,
consumer sync, planning.
Handoff ready.
Next session cold-starts cleanly.
Works with any language, any AI tool, any CI system.
| Command | What it does |
|---|---|
| Onboarding | |
| trace init | New project — full enforcement from day one |
| trace scan | Existing project — auto-detection + baseline |
| Coherence | |
| trace check | Full 5-pillar coherence validation |
| trace status | Quick health overview |
| trace deps <id> | Dependency graph + transitive impact analysis |
| trace impact <id> | Pre-work blast radius + consumer sync snapshot |
| trace search | Full-text search across all TRACE artifacts |
| Session | |
| trace gate start | Integrity + config validation + AI context (supports --scope) |
| trace gate end | Quality tools + consumer sync + planning + metrics |
| trace checkpoint | Crash recovery + auto-observation via git diff |
| trace override | Emergency bypass with tracked debt |
| Planning | |
| trace plan | Kanban-style backlog board in YAML |
| trace plan add | Add item with priority and sprint |
| trace plan move | Move item between status columns |
| trace plan release | Auto-generate release notes from completed items |
| CI/CD & Maintenance | |
| trace ci | PR-scoped analysis with JSON + GitHub PR commenting |
| trace metrics | Outcome analysis + threshold calibration |
| trace integrity | SHA-256 tamper detection for TRACE files |
| trace validate | Config validation with "did you mean?" suggestions |
| trace upgrade | Schema migration — add missing sections safely |
| trace watch | File monitor + auto-session when AI skips gates |
| trace license | Scan dependencies for license compliance |
| trace deps audit | Dependency governance (policy, security, audit) |
| AI Process Enforcement | |
| trace hook install | Pre-commit hook — blocks incoherent commits |
| trace hook uninstall | Remove the pre-commit hook |
| trace hook status | Check if hook is active |
| MCP Server | |
| trace mcp | Start MCP server (AI calls TRACE automatically) |
| trace mcp setup | Show config for Claude Code, Cursor, Kiro |
Reads file structure, not file content. Nothing phones home.
Connect the MCP server to your AI tool. TRACE checks impact before changes, validates coherence after, and audits dependencies automatically. You never type trace gate start again.
trace mcp exposes 6 tools to your AI. It reads context, checks impact, validates coherence, and audits deps — all automatically.
AI_INSTRUCTIONS.md tells AI to follow gates with mandatory gate statements. Behavioral nudge for tools without MCP.
trace watch detects ungated file changes, auto-opens sessions, logs to PROJECT_LOG. History never lost.
trace hook install blocks incoherent commits. The hard stop that catches everything else.
trace scan and you're started. Auto-detects anchors, consumers, tests, quality tools. Uses "Clean as You Code" so only new code is enforced.trace watch captures changes even when AI skips gates, and trace hook install blocks incoherent commits. No single layer is enough. All three together close the gap.trace deps audit checks all packages against configurable policy rules: blocked packages, license restrictions, staleness, and security vulnerabilities. Three modes: strict (allowlist only), moderate (rule-based), and permissive (flag only). Runs automatically at gate end.yaml package. Everything else is Node.js core. 53 KB compressed.Deep dives, real-world stories, and practical guides.
After 40+ sessions, I found three failure patterns in every AI-assisted project. Here's how TRACE solves each one.
Read articleThe invisible failure mode of AI-assisted development, and why structural coherence is the fix.
Read articleInstall, scan your project, run your first gate. A step-by-step walkthrough with real output.
Read articleAI systems are stateless. Human memory is unreliable. TRACE creates persistent structural memory for your codebase.