ladder-drift-and-meta-aware-regression

ladder-drift-and-meta-aware-regression is a skill for Claude Code, Codex from topprismdata/cultivating-ml-agent. It costs 178 tokens per session (2,232 once invoked), scanned A, original, MIT.

A set of lessons for diagnosing changing results in an AI card-game agent, including leaderboard changes caused by shifts in the opponents' popular strategies.

In plain words
What is it for?
Use it to investigate sudden leaderboard drops, inspect changes in the current game meta, evaluate strategy adjustments, and compare agent variants with larger samples.
Why use it?
It helps distinguish a real code regression from a changing evaluation environment and warns against trusting results from very small tests.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Good fit Use it to investigate sudden leaderboard drops, inspect changes in the current game meta, evaluate strategy adjustments, and compare agent variants with larger samples.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/topprismdata/cultivating-ml-agent/ladder-drift-meta-aware-regression
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add topprismdata/cultivating-ml-agent --skill ladder-drift-meta-aware-regression
Clone the repo
git clone --depth 1 https://github.com/topprismdata/cultivating-ml-agent

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for ladder-drift-and-meta-aware-regression

README.md
[![agentmods](https://agentmods.dev/badge/skills/topprismdata/cultivating-ml-agent/ladder-drift-meta-aware-regression/github.svg)](https://agentmods.dev/skills/topprismdata/cultivating-ml-agent/ladder-drift-meta-aware-regression)
Your own site
<a href="https://agentmods.dev/skills/topprismdata/cultivating-ml-agent/ladder-drift-meta-aware-regression"><img src="https://agentmods.dev/badge/skills/topprismdata/cultivating-ml-agent/ladder-drift-meta-aware-regression/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for ladder-drift-and-meta-aware-regression

Your own site · 80×15
<a href="https://agentmods.dev/skills/topprismdata/cultivating-ml-agent/ladder-drift-meta-aware-regression"><img src="https://agentmods.dev/badge/skills/topprismdata/cultivating-ml-agent/ladder-drift-meta-aware-regression.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 178 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,232 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00178 $0.02232
Opus 5 $0.00089 $0.01116
Sonnet 5 $0.00036 $0.00446
Haiku 4.5 $0.00018 $0.00223

Measured 11d ago against content hash f57460519ca3, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

ladder-drift-and-meta-aware-regression scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/examples/ladder-drift-meta-aware-regression/SKILL.md · 167 lines

How it starts

The opening of the file, as written. The whole thing — 167 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Ladder Drift + Meta-Aware Regression: The PTCG AI Battle 2026-06 Findings

This skill captures three hard-won lessons from a multi-week PTCG (Pokémon TCG) AI Battle project where a strong 1084.5-LB Lucario agent (v29, public notebook fork) held the lead for days — then lost 200 LB in 48 hours due to meta shift, and lost another 150 LB to "meta-aware" patches that LOOKED like improvements in 60-game tests but REGRESSED in 200-game tests.

Lesson 1: Ladder Drift is Real (Same Code, -200 LB in 48h)

v29 baseline (no code changes) — different scores across days:

Date LB Score Delta
2026-06-27 13:13 970.0 baseline
2026-06-28 16:29 774.6 -195
2026-06-29 16:25 770.5 -4 (stabilizing)

The agent didn't change. The ladder did. Top 5 decks shifted from Lucario-dominant to Starmie / Archaludon / Dragapult-dominant. The same code that scored 970 against a Lucario ladder scored 770 against a Starmie ladder.

Implication: When your agent's LB drops, check the meta before debugging the code. Look at top-20 leaderboard deck names; if they're all different decks than when you calibrated, it's drift, not bug.

Detection: Compare LB score to baseline over time. If drift > 100 points in a week with no code change, the meta moved. Don't waste quota on resubmits of the same code.

Lesson 2: Meta-Aware Heuristics Often REGRESS (The 60→200 Game Lesson)

Three "improvements" tested on small samples (60-80 games) all looked neutral or marginally positive. All three regressed when tested at 200 games:

Variant Sample Mirror vs Meta Deck Verdict (60g) Verdict (200g)
v33 (v29 + ALL v31 logic) 80g mirror + 60g meta -7.5% mirror -13.1% vs Arch "neutral" REGRESSION
v34 (v29 + Archaludon-only detection) 80g mirror + 60g meta +5.0% mirror -5.6% vs Arch "slightly better" -10.1% vs Arch
v34 v2 (rerun) 200g mirror + 200g meta +10.0% mirror -25.9% vs Arch n/a -10.1% vs Arch

Read the full file on GitHub · 167 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 11d ago First seen · 167 lines · 178 tokens per session scan A f57460519ca3

Subscribe to this mod's changes

ladder-drift-and-meta-aware-regression is a skill published in the GitHub repository topprismdata/cultivating-ml-agent (5 stars, last pushed 14d ago), licensed MIT. It adds 178 tokens to every session and 2,232 once invoked, about $0.0009 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

huggingface-hub

Hugging Face Hub CLI (hf) — search, download, and upload models and datasets, manage repos, query datasets with SQL, deploy inference endpoints, manage Spaces and buckets.

braxtonROSE4/zorro-agent · 43 tokens

tensorboard

Visualize training metrics, debug models with histograms, compare experiments, visualize model graphs, and profile performance with TensorBoard - Google's ML visualization toolkit.

davila7/claude-code-templates · 32 tokens

mlflow

Track ML experiments, manage model registry with versioning, deploy models to production, and reproduce experiments with MLflow - framework-agnostic ML lifecycle platform.

davila7/claude-code-templates · 33 tokens

datachain-knowledge

Use whenever datasets, cloud storage buckets, or data pipelines are mentioned — creating, saving, querying, listing, exploring, deleting, or processing data in S3, GCS, Azure Blob, or local storage. Also use when running any script that may create datasets as a side effect. Maintains a knowledge base at dc-knowledge/…

datachain-ai/datachain · 104 tokens

prompt-scanner

A scanner for text sent to an AI agent, looking for prompt injection and jailbreak attempts. Prompt injection is text that tries to override an agent's instructions; a jailbreak tries to bypass its safety limits.

alibaba/anolisa · 103 tokens

install-openviking-memory

Install and configure the OpenViking long-term memory plugin for OpenClaw via natural conversation. Once installed, the plugin automatically captures facts from chats and recalls relevant context before each reply (auto-capture + auto-recall, cross-session). Covers prerequisites, install through OpenClaw's plugin…

volcengine/OpenViking · 191 tokens