Works with the AI Agents you already use
Model Migration Planner
Move a workload to another model with a parity matrix, cost and latency projections, and a staged rollout you can abort.
$15
Model Migration Planner
Example session with this skill installed
We are on a frontier model and the bill is too high. We want to move summarisation and ticket triage to a cheaper one next month. What breaks and how do we roll it out?
- Read your context and instructions
- Compiled the model migration planner
Model migration: current frontier model -> cheaper target
Recommendation
Go, with conditions. Cost $2,100 -> $540/month at current volume.
Quality on 80 real cases: summarisation 91% -> 89% (within noise), triage 88% -> 74% (blocking).
Move summarisation only; keep triage where it is.
Parity matrix
| Capability | Current | Target | Gap |
|---|---|---|---|
| Context window | 200k | 128k | 9 inputs exceed it |
| Structured output | Schema mode | Loose JSON | 6% invalid |
| Tool calling | Parallel | Sequential | Slower multi-tool flows |
| p95 latency | 2.1s | 1.4s | None |
| Price in/out | 3.00 / 15.00 | 0.25 / 1.25 | 88% cheaper |
Prompt changes
Remove three "do not" instructions the old model needed; re-test formatting; cap max tokens at 800 for summaries.
Rollout
Shadow 100% for one week, canary 5%, ramp 25/50/100.
Gate: quality at or above 87%, error rate below 1%.
Abort trigger: quality below 85% or any rise in invalid JSON.
Rollback: old path stays deployable for two weeks.
Unknowns
Triage quality gap is unexplained; investigate before any second attempt.
Connects securely to your tools. The creator never sees your data.
About this skill
Model migration plan: parity matrix on your own cases, cost and latency projections, and a shadow-canary-ramp rollout with an abort trigger.
Use this skill when:
- switch models for cost or quality
- check if a cheaper model is good enough
- a model is being deprecated
- plan a staged model rollout
Who it's for: Teams changing models without breaking quality.
Example request
Move ticket triage to a small model next month. What breaks?
What the example returns
- Triage parity 88% to 74%: blocking
- Summaries within noise, down 2
- Shadow, canary, gate at 85%
In the package: five files — SKILL.md with the trigger and workflow, README.md, references/METHOD.md (the full method), references/CHECKLIST.md (the pass/fail gate) and examples/EXAMPLE.md (a worked run you can copy).
Works with Claude Code, Cursor, Codex CLI and any agent that reads SKILL.md. Plain markdown: no API key, no account, no network access.

Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for LLM Ops.
Requires an existing evaluation set or production logs. Does not handle API key management or CI/CD deployment execution.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 2 days ago