drift-sense

Detect deviations across codebase, canonical memory, and traces with YAML findings.

2|Updated Apr 25, 2026
One-click install
npx skills add https://github.com/iurykrieger/claude-yoke --skill drift-sense
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: drift-sense
Source: https://github.com/iurykrieger/claude-yoke/tree/main/skills/drift-sense
Command: npx skills add https://github.com/iurykrieger/claude-yoke --skill drift-sense

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Drift-sense provides continuous monitoring to detect deviations across the host codebase, the canonical-memory layer, and historical traces, enabling proactive governance of AI-driven processes.

Core Features & Use Cases

  • Three-mode drift analysis: codebase, canonical-memory, and traces, with structured YAML findings.
  • Early-deprecation signals: surface model drift, memory misalignment, and recurring contract issues for canonization review.
  • Use Case: For a running Claude-based agent, run drift-sense to surface drift signals before deployment and review the YAML report to decide mitigations.

Quick Start

Invoke drift-sense to scan codebase, canonical memory, and traces and review the generated YAML findings.

Frequently Asked Questions about drift-sense

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I detect memory drift and codebase deviations in AI agent workflows?

Drift monitoring identifies memory drift and codebase deviations by comparing current states against historical traces. It emits structured YAML findings to flag misalignment signals for proactive AI governance review before deployment.

What is canonical memory drift and when do I need to monitor it?

Canonical memory drift occurs when AI agent memory diverges from established codebase contracts over time. You need to monitor it in production AI governance workflows to catch early-deprecation signals and recurring contract issues before they cause failures.

How do I run trace analysis to surface drift signals across a production codebase?

Run trace analysis by executing drift-sense with the traces target mode. It scans historical execution traces alongside the codebase and outputs structured YAML findings that highlight deviations requiring canonization review.

Can I use drift-sense with GitHub Actions for continuous codebase monitoring?

Yes, drift-sense integrates with GitHub Actions through the workflow file at .github/workflows/yoke-drift-sense.yml, enabling continuous drift monitoring across codebase, canonical memory, and traces within your CI/CD pipeline.

Does drift-sense require external dependencies to analyze canonical memory?

No, drift-sense operates with zero external dependencies. It analyzes canonical memory, codebase files, and traces natively and emits structured findings directly in YAML format without requiring additional packages.

What is the best way to format AI governance drift findings for automated review?

The best way to format AI governance drift findings is using YAML reports. Drift-sense emits structured YAML findings across codebase, canonical-memory, and trace modes, making outputs machine-readable for automated canonization decisions.