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

76%

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

Fix and improve this skill with Tessl

tessl review fix ./skills/mle-workflow/SKILL.md

Quality

Content

63%Weight 40%Scale 1-5

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

The body is a thorough, well-structured ML engineering workflow with executable code, templates, and a clear six-step sequence with validation gates. Its main weaknesses are conciseness (large inlined tables) and progressive disclosure (everything is monolithic with no external references).

Suggestions

Move the exhaustive SWE-surface mapping table and/or the 10 MLE task simulations into a separate reference file, keeping SKILL.md a concise overview with one-level-deep links, to improve progressive disclosure.

Tighten the enumerative lists (e.g., Scope Calibration, Metric and Mistake Economics) into the highest-signal items to reduce token bulk without losing the core workflow.

Add an explicit runnable validate->fix->retry loop (e.g., a promotion-gate CLI invocation with retry on failure) for the deploy/rollback steps to strengthen workflow_clarity feedback loops.

Dimension	Reasoning	Score
Conciseness	Mostly efficient and free of explanations of concepts Claude already knows, but the large SWE-surface table (20+ rows) and 10-task simulation table add enumerative bulk that could be tightened.	3 / 5
Actionability	Provides executable Python (TrainingConfig, artifact_name, PROMOTION_GATES, assert_promotion_ready) plus concrete templates (Iteration Compact, Observation Ledger) and a review checklist, though much guidance is principle-list style rather than runnable commands.	4 / 5
Workflow Clarity	A clearly sequenced 6-step Core Workflow with named checkpoints (promotion gates that fail closed, 'do not ship' thresholds, rollback plans), though explicit validate->fix->retry command loops for batch/deploy operations are described rather than scripted.	4 / 5
Progressive Disclosure	Well-organized with clear section headers, but no bundle files exist and content that could live in separate references (the exhaustive SWE-surface mapping and the 10 MLE task simulations) is fully inlined in one ~340-line SKILL.md.	3 / 5
	Total	14 / 20 Passed

Description

88%Weight 40%Scale 1-5

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 concrete and complete, clearly stating both the workflow's capabilities and a natural 'Use when' trigger. It is well-scoped to production ML with minimal conflict risk, though a few additional common synonyms would round out trigger coverage.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'data contracts, reproducible training, model evaluation, deployment, monitoring, and rollback' — giving comprehensive coverage of the workflow's capabilities.	5 / 5
Completeness	Explicitly answers both what ('Production machine-learning engineering workflow for ...') and when ('Use when building, reviewing, or hardening ML systems beyond one-off notebooks.') with concrete trigger phrases.	5 / 5
Trigger Term Quality	Natural phrases like 'building, reviewing, or hardening ML systems' and 'beyond one-off notebooks' are present, but a few common variations (e.g., 'MLOps', 'model training', 'model refresh') users might say are missing.	4 / 5
Distinctiveness Conflict Risk	Has a clear niche (production ML engineering, scoped beyond one-off notebooks) with distinct triggers, but minor overlap risk with related SWE/deployment skills.	4 / 5
	Total	18 / 20 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
metadata_version	'metadata.version' is missing	Warning

	Total	15 / 16 Passed

Repository: affaan-m/ECC
Path: skills/mle-workflow/SKILL.md
Commit: d8409a4

Reviewed: 1 day 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.