Name: mle-workflow
Rating: 72 (1 reviews)
Author: affaan-m

mle-workflow

Production machine-learning engineering workflow for data contracts, reproducible training, model evaluation, deployment, monitoring, and rollback. Use when building, reviewing, or hardening ML systems beyond one-off notebooks.

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

The canonical home for this skill is mle-workflow in affaan-m/ECC

Quality

Content

77%Weight 40%Scale 1-3

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is highly actionable with executable code, templates, and a well-sequenced workflow with validation checkpoints, but it is over-long for a single file and retains conceptual primer content Claude already knows. Splitting the large reference tables into bundle files would improve both conciseness and progressive disclosure.

Suggestions

Trim the "Metric and Mistake Economics" and "Data and Feature Hypotheses" primers to the non-obvious workflow guidance; assume Claude knows precision/recall tradeoffs and missing-value/outlier basics.

Move the 20-row SWE surface table and 10-task simulation matrix into reference files (e.g. references/swe-surface.md, references/task-simulations.md) and link to them so SKILL.md stays an overview.

Add one-line in-page pointers above the large tables so the overview can be scanned before drilling into the detail.

Dimension	Reasoning	Score
Conciseness	Mostly actionable but sections like "Metric and Mistake Economics" and "Data and Feature Hypotheses" restate basic ML concepts (precision/recall tradeoffs, missing-value and outlier handling) Claude already knows; the body runs ~340 lines. Not (3) because several tokens do not earn their place; not (1) because the core is efficient and non-generic.	2 / 3
Actionability	Provides executable code (TrainingConfig dataclass, artifact_name, fail-closed promotion-gate asserts), concrete templates (Iteration Compact, Observation Ledger), and task-to-tool mapping tables — copy-paste ready rather than pseudocode.	3 / 3
Workflow Clarity	The six-step Core Workflow is clearly sequenced with explicit validation checkpoints (fail-closed promotion gates, leakage checks), plus a review checklist and numbered Decision Brain and Error Analysis loops. Not (2) because checkpoints are explicit, not implicit.	3 / 3
Progressive Disclosure	Well-organized into sections but all ~340 lines are inline in SKILL.md with no bundle files; the 20-row SWE surface table and 10-task simulation matrix are reference material that should be split out. Not (3) because nothing is offloaded to one-level-deep references; not (1) because sections are clearly labeled rather than a monolithic wall.	2 / 3
	Total	10 / 12 Passed

Description

100%Weight 40%Scale 1-3

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is strong: it names concrete capabilities, supplies an explicit "Use when" trigger with natural phrasing, and carves out a distinct production-ML niche. No significant gaps relative to the rubric anchors.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — "data contracts, reproducible training, model evaluation, deployment, monitoring, and rollback" — matching the multi-action anchor rather than the single-action (2) or vague (1) level.	3 / 3
Completeness	Explicitly answers both what (the workflow and its phases) and when via "Use when building, reviewing, or hardening ML systems beyond one-off notebooks", matching the explicit-trigger anchor.	3 / 3
Trigger Term Quality	Includes natural trigger terms a user would say — "building, reviewing, or hardening ML systems" and "beyond one-off notebooks" — with good coverage; not the partial (2) level since common phrasings are present.	3 / 3
Distinctiveness Conflict Risk	Targets a clear MLE production niche with distinct triggers unlikely to overlap with general SWE or notebook skills; not the (2) level because the scope is sharply bounded to production ML.	3 / 3
	Total	12 / 12 Passed

Validation

93%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: affaan-m/ECC
Path: .kiro/skills/mle-workflow/SKILL.md
Commit: 0c1d7be

Reviewed: 3 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.