data-leakage-hunting SKILL.md skillA

data-leakage-hunting from annatchijova/skills is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.

Install it, pinned
mdr add annatchijova/skills/data-leakage-hunting@git:20260822.33d806d

Pins git:20260822.33d806d in mdr.lock. Pin the fingerprint instead.

Audit grade
A, 17 of 17 checks passed
Kind
SKILL.md skill
Repository
annatchijova/skills
Path
data-leakage-hunting/SKILL.md
Versions recorded
1
Latest version
git:20260822.33d806d, committed Aug 22, 2026
Size
10,527 bytes
Repository stars
9
License
a license file GitHub does not identify

Hunt the one bug class in machine learning whose symptom is a better score — information from the evaluation set reaching the model through preprocessing, time, groups, duplicates, features derived from the label, or a test set spent on tuning. Use whenever a dataset is split, a pipeline is fitted, a feature is engineered, a benchmark is reported, or a model performs suspiciously well. Trigger on "data leakage", "target leakage", "train/test contamination", "the accuracy is 99%…

Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.

What the file says

# Data Leakage Hunting

Every other defect makes something fail. This one makes everything look
better — which means the normal feedback loop of engineering runs in reverse.
A leak is discovered by celebration, adopted into the pipeline, published in
the README, and found six months later by production.

So the alarm has to be inverted, and it is the first rule of this skill:

> **An unexpectedly good result is a bug report until proven otherwise.**

The invariant being violated is always the same one, and it is worth writing on
the wall:

> **No information derived from the evaluation data may influence anything used
> to produce the prediction.** Established at the split (T0), it must still hold
> at fit time, at feature-construction time, at tuning time, and at the moment
> the number goes in the report (Tn).

That is `invariant-hunting`'s shape exactly: a property established in one place
that has to survive a journey. The difference is that here the violation
*rewards* you, so nobody goes looking.

Composes with the library:

- **invariant-hunting** — the T0→Tn frame; this is its ML instance
- **model-evaluation-discipline** — the number the leak corrupts, and the
…

Read the whole file at its exact version.

How to install

The exact contents, by fingerprint
mdr add annatchijova/skills/data-leakage-hunting@sha256:b85023097923ddb3

The line at the top of this page pins the label git:20260822.33d806d, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.

Badge

mdr badge

[![mdr](https://markdownregistry.com/badge/art_wes6dhy45aufwro3.svg)](https://markdownregistry.com/a/art_wes6dhy45aufwro3)

2 badge image fetches in 30 days. What the badge shows.

Versions

The registry has recorded 1 version of this file, committed on Aug 22, 2026.

versioncommittedcommitsizeaudit
git:20260822.33d806d latest2026-08-22 33d806d 10,527 BA view

Audit of the latest version

A  17 of 17 checks passed. Fixed automatic checks with no AI involved, so the same checks give the same answer every run.
  • pass: Frontmatter block present
  • pass: Frontmatter declares a name
  • pass: Frontmatter declares a description
  • pass: Size between 200 bytes and 200 KB (10527 bytes)
  • pass: No zero-width or bidi control characters
  • pass: No instruction hidden inside an HTML comment
  • pass: No link to an exfiltration or paste host
  • pass: No credential-shaped string
  • pass: No instruction to send local credentials anywhere
  • pass: No text hidden with inline styles
  • pass: No prompt-injection phrasing
  • pass: No curl or wget piped into a shell
  • pass: No recursive delete of root, home or parent
  • pass: No instruction to read or print local credentials
  • pass: No base64 blob over 200 characters
  • pass: No link to a raw IP address
  • pass: No script tag

Source

GitHub

annatchijova/skills · 9 stars · license a license file GitHub does not identify · pushed 2026-11-02 · branch main

API

GET https://markdownregistry.com/api/v1/artifacts/art_wes6dhy45aufwro3
GET https://markdownregistry.com/api/v1/resolve?ref=annatchijova/skills/data-leakage-hunting
GET https://markdownregistry.com/api/v1/blob/b85023097923ddb3f52f089243d3700e78eefa4e86e7dfd87fcf5131657012c2

Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.

More from annatchijova/skills

abductive-engineering SKILL.md skill
annatchijova/skills · abductive-engineering/SKILL.md · Software engineering best practices grounded in C.S. Peirce's abductive reasoning and his triadic framework…
git:20260708.f31c428 · audit A · 9 stars
acquisition-order-of-volatility SKILL.md skill
annatchijova/skills · acquisition-order-of-volatility/SKILL.md · Collect digital evidence in order of volatility, or destroy the thing you came for — memory, network state, and running…
git:20260825.8500180 · audit A · 9 stars
agent-trust-boundaries SKILL.md skill
annatchijova/skills · agent-trust-boundaries/SKILL.md · Design LLM and agentic systems so that retrieved content is data and never instruction — instruction authority comes…
git:20260815.2390440 · audit A · 9 stars
alert-triage-economics SKILL.md skill
annatchijova/skills · alert-triage-economics/SKILL.md · Triage a security alert queue by expected loss, not by the severity label the tool stamped — because analyst attention…
git:20260825.8500180 · audit A · 9 stars
assume-breach-modeling SKILL.md skill
annatchijova/skills · assume-breach-modeling/SKILL.md · Start from "this step already succeeded" and map what the attacker reaches next — the identity held at that position…
git:20260822.f25b32b · audit A · 9 stars
atomic-state-mutation SKILL.md skill
annatchijova/skills · atomic-state-mutation/SKILL.md · Make a logical operation that spans several writes to persistent state land all-or-nothing, isolated from concurrent…
git:20260708.f31c428 · audit A · 9 stars
attack-surface-triage SKILL.md skill
annatchijova/skills · attack-surface-triage/SKILL.md · Enumerate the attack surface of an AUTHORIZED target and rank it into a reproducible, falsifiable candidate queue —…
git:20260815.2390440 · audit A · 9 stars
audit-before-patch SKILL.md skill
annatchijova/skills · audit-before-patch/SKILL.md · Validate any audit finding, bug report, or proposed fix against the actual current file content before changing a…
git:20260708.f31c428 · audit A · 9 stars
authorization-surface-mapping SKILL.md skill
annatchijova/skills · authorization-surface-mapping/SKILL.md · Build the actor × resource × action matrix a system implicitly claims to enforce, then test the cells it never wrote a…
git:20260822.f25b32b · audit A · 9 stars
beyond-the-fix SKILL.md skill
annatchijova/skills · beyond-the-fix/SKILL.md · Audit the bug fixes themselves — a patch proves that someone knew something was wrong there, never that it is now…
git:20260822.9eda33b · audit A · 9 stars
beyond-the-sink SKILL.md skill
annatchijova/skills · beyond-the-sink/SKILL.md · Look past the obvious layer of an investigation — past sink-grep keyword lists, past the exhausted question family…
git:20260821.38a2b1d · audit A · 9 stars
claim-provenance-discipline SKILL.md skill
annatchijova/skills · claim-provenance-discipline/SKILL.md · Tracking and preserving the origin of every assertion as it travels through drafts, handoffs, summaries, and documents…
git:20260710.eb435f4 · audit A · 9 stars

Every file in annatchijova/skills

Browse by kind, by grade A, or by owner.