model-migration

Source code for the paper "Accelerating Hierarchical Federated Learning under Mobility via Model Migration in Cloud–Edge–End Collaborative Networks".

Open-source LLM migration and regression testing for AI agents. Compare models, detect tool-call regressions, and gate model changes in CI.

Migrate LLM apps between models safely — then benchmark old vs. new to prove the upgrade. Agent skill for behavior-preserving migrations + a runnable eval harness (three-arm leaderboard, prompt sweeps). Works in Claude Code, Codex & Cursor.

Early-stop canary testing for LLM model migrations

A portable memory skill and CLI for AI agents: curated context, user preferences, topic interruptions, handoff packets, and model/runtime migration without transcript dumps.

Local, cross-provider preflight checks for LLM integration changes

Know what breaks before switching LLMs.

Why a passing benchmark isn't safe to ship: a free 2-stage (benchmark + replay) validation recipe for LLM model swaps & prompt changes, run on flat-rate coding-agent subagents — no eval API bill.

Plan, research, and evaluate LLM application migrations between models, providers, and platforms (Claude, GPT, Bedrock) before changing code

Dependabot for AI models — replay your app's scenarios across a model migration and catch real behavioral regressions before they ship. Low false-positive by design. CLI + GitHub Action. Open-core, Apache-2.0.

Ollama model store porter: re-hash every blob, export/import models offline as verified tarballs

git diff for LLM behaviour. Capture what your model actually did, replay it against a candidate model, and catch what silently broke. No test authoring, no API keys required.

Portable, content-addressed reliability evidence for LLM systems. Capture how a model behaves under perturbation; preserve, verify, and diff the evidence across model changes.

Audit GPT-6.1 Sol API requests for endpoint, tool, reasoning, residency, context and token-cost compatibility before migration.

Agent Skill — single entry for auditing self-authored Claude Code assets when a new Claude model takes a role: a runtime cross-check against the loaded system prompt plus a dated-pattern scan via /claude-api prompt-audit; applies safe line fixes and hands verdicts to the stocktake skills. Implements AKC Scaffold Dissolution's generation trigger.

Independent, reproducible DeepSeek V4.1 Flash migration evaluation and decision service

[ARCHIVED 2026-09] An experiment in detecting and repairing AI-agent regressions across model upgrades. Development paused. Strongest result: 36/38 regressions found, 32 repaired, 4 not. MIT, runs locally.

GitHub Action for LLM migration and regression testing — runs your golden suite on every PR and fails the check on model regressions.

A verified production LLM model migration, and how its method evolved into ModelPromote: blind evaluation, human approval, read-back activation, telemetry verification and rollback. Runnable offline on synthetic data.

Govern the switch when you change the AI model in production: require approval, activate with read-back confirmation, verify from telemetry, and roll back to a locked target. Leaves one portable migration record. TypeScript, zero runtime dependencies.