π Model Migration Checker
Find the model-version coupling that breaks when you swap LLMs. Flags hardcoded model names and versions, deprecated or renamed parameters (the max_tokens to max_completion_tokens class of change), hardcoded token and context-window limits, response-format parsing tied to one model's output, tool-schema format coupling between providers, and hardcoded per-token cost constants. The patterns load from an editable model-rules table you update as new models ship.
- Scan codebases for hardcoded LLM model names and versions
- Identify deprecated API parameters and legacy token limit assumptions
- Detect rigid response-parsing logic that risks breaking after a model swap
$15
Β· or 75 creditsSecure checkout via Stripe
Included in download
- Scan codebases for hardcoded LLM model names and versions
- Identify deprecated API parameters and legacy token limit assumptions
- terminal, file_read automation included
- Ready for Cursor
Sample input
Scan my project for hardcoded model versions or legacy parameters before I migrate to the new Claude 3.5 Sonnet.
Sample output
Migration Audit Report
- MMC001 (High): Found 'gpt-4-0613' in
services/openai_client.py:42 - MMC003 (Med): Hardcoded 4096 token limit in
utils/tokenizer.ts:12 - MMC006 (Low): Pricing constant ($0.03/1k) detected in
billing/calculator.js:8 - Recommendation: Abstract model strings to env vars.
π Model Migration Checker
Find the model-version coupling that breaks when you swap LLMs. Flags hardcoded model names and versions, deprecated or renamed parameters (the max_tokens to max_completion_tokens class of change), hardcoded token and context-window limits, response-format parsing tied to one model's output, tool-schema format coupling between providers, and hardcoded per-token cost constants. The patterns load from an editable model-rules table you update as new models ship.
$15
Β· or 75 creditsSecure checkout via Stripe
Included in download
- Scan codebases for hardcoded LLM model names and versions
- Identify deprecated API parameters and legacy token limit assumptions
- terminal, file_read automation included
- Ready for Cursor
- Instant install
Sample input
Scan my project for hardcoded model versions or legacy parameters before I migrate to the new Claude 3.5 Sonnet.
Sample output
Migration Audit Report
- MMC001 (High): Found 'gpt-4-0613' in
services/openai_client.py:42 - MMC003 (Med): Hardcoded 4096 token limit in
utils/tokenizer.ts:12 - MMC006 (Low): Pricing constant ($0.03/1k) detected in
billing/calculator.js:8 - Recommendation: Abstract model strings to env vars.
About This Skill
Automate LLM Model Migrations
Upgrading your codebase to a newer LLM model (e.g., GPT-4 to GPT-4o, or Claude 3 to 3.5) often breaks logic due to hardcoded version strings, deprecated API parameters, or legacy token limits. The Model Migration Checker is a specialized audit tool designed for developers who need to identify migration risks before they hit production errors.
What it does
This skill scans your repository (Python, JavaScript, TypeScript, and config files) to detect six critical categories of model coupling:
- Hardcoded Models: Specific version strings nested in API calls.
- Deprecated Parameters: Arguments no longer supported by modern provider SDKs.
- Context Assumptions: Legacy token limits or hardcoded context window sizes.
- Coupling: Rigid response-parsing logic or function-calling schemas that break with model output variations.
- Cost Constants: Hardcoded pricing data used for internal usage tracking.
Why use this skill
Unlike a generic AI prompt, this skill uses a structured heuristic engine to catch "silent failures" that a human or a standard LLM call might overlook. It provides a technical audit trail with specific rule IDs and severity levels, allowing you to systematically clear blockers before a model swap. It ensures your infrastructure is model-agnostic and resilient to future provider updates.
Use Cases
- Scan codebases for hardcoded LLM model names and versions
- Identify deprecated API parameters and legacy token limit assumptions
- Detect rigid response-parsing logic that risks breaking after a model swap
- Find hardcoded per-token pricing constants in billing modules
Known Limitations
Heuristic detector. It flags coupling to review before and after a swap; it does not run your code or talk to a model API, and the rules table is as current as you keep it.
How to Install
mkdir -p ~/.claude/skills && curl -sL https://www.agensi.io/api/install/model-migration-checker -o /tmp/model-migration-checker.zip && unzip -o /tmp/model-migration-checker.zip -d ~/.claude/skills && rm /tmp/model-migration-checker.zipFree skills install directly. Paid skills require purchase - use the download button above after buying.
Reviews
No reviews yet - be the first to share your experience.
Only users who have downloaded or purchased this skill can leave a review.
Early access skill
Be the first to review this skill.
Only users who have downloaded or purchased this skill can leave a review.
Security Scanned
Passed automated security review
Permissions
File Scopes
Match patterns (model-name shapes, deprecated params, known limits) load from an editable references/model-rules.json, so you can update them the moment a new model lands. Reads no environment variables.
Works with any agent that can read a repo and run a local Python script (Claude Code, Cursor, Codex CLI, and other SKILL.md-compatible agents). Standard library only, no install step. Read-only, no network.