Data Science Engineering Foundation — Agent Skill by Microsoft

From the author

State, resume, reconstruction, job-lifecycle, transition, and flow-state mechanics for the Data Science and Engineering Coach. Loaded by the coach; not a user entry point.

— the description by Microsoft

AutoGineer summary

Written by AutoGineer from the skill’s own files, not by the author

The Data Science and Engineering Foundation serves as the core orchestration layer for the Data Science and Engineering Coach. It provides a centralized set of protocols for managing session state, job lifecycles, and flow-state mechanics. This foundation ensures consistency across different coaching sessions and job transitions by defining how the coach should initialize, resume, and navigate between various tasks. It acts as an internal reference index rather than a user-facing tool, enforcing rules for when to interrupt, how to handle durable writes, and how to maintain session integrity.

What it does

Defines protocols for session initialization, state reconstruction, and resumption.
Provides a registry for routing tasks to specific skills and specialists.
Implements a flow-state protocol to manage interruptions and user gates.
Enforces a durable-write gate to scan for sensitive content before creating artifacts.
Maintains clear boundaries between general orchestration and job-specific logic.
Standardizes the lifecycle management of coaching jobs.

When to use it

When the coach needs to initialize or resume a session.
When transitioning between different data science or engineering jobs.
Before performing a durable write of a customer artifact.
When validating the state of a coaching session.
When determining the appropriate skill or specialist to handle a specific task.

Advantages

Ensures consistent behavior across different coaching sessions.
Provides a structured approach to handling sensitive data during artifact creation.
Decouples orchestration logic from job-specific implementation details.
Improves reliability by enforcing clear state and lifecycle rules.
Facilitates modularity by routing tasks to specialized skills.

Usage examples

“Initialize the coaching session for a new data engineering project.”: The coach loads the session-state protocol, validates the project identity, and prepares the environment for the selected job.
“I want to save this new data catalog document.”: The coach triggers the flow-state protocol, performs a sensitive content scan on the document, and proceeds with the write only if the scan passes.
“Switch from the current feasibility study to a new pipeline design task.”: The coach follows the transition protocol to pause the current job, confirm the switch with the user, and route to the dataops skill.

Install

Download the skill as a ZIP:
https://mp-staging.autogineer.com/agent-skills/microsoft/data-science-engineering-foundation/download

Unzip it into the cross-client skills folder of your project:

mkdir -p .agents/skills && unzip data-science-engineering-foundation.zip -d .agents/skills

Documentation by Microsoft

Data Science and Engineering Foundation

Goal

Keep Data Science and Engineering Coach orchestration consistent across job changes and sessions without duplicating job-specific guidance. The coach loads this index at initialization and resume, then reads the reference for the current coaching moment.

Reference index

Reference	When to read
job-registry.md	Before offering or selecting a job, routing to a skill or specialist, or naming a durable output
lifecycle-classes.md	When starting, pausing, resuming, completing, or re-invoking a job
transition-protocol.md	When a topic shift, explicit request, or completion suggests moving between jobs
session-state.md	Before initialization, validation, mutation, resume, recovery, or reconstruction of coaching state
flow-state.md	Before interrupting work, crossing a gate, writing a durable artifact, or offering post-job choices

Success criteria

Each orchestration rule has one owner in this package.
State and lifecycle mechanics remain independent of job-specific methods.
The coach reads the applicable reference before acting on its contract.

Constraints

Keep catalog guidance in data-catalog, the durable data-catalog workflow for entities, declared relationships, lineage, and ERD-ready model semantics.
Keep feasibility guidance in feasibility, the evidence-led data and ML feasibility-study workflow with lifecycle and interchange traceability.
Keep pipeline and testing guidance in dataops, the DataOps reference for tier behavior, pipeline invariants, validation placement, tests, drift, and operational signals.
Keep general experiment guidance in experiment-design, the reusable workflow for candidate selection, hypotheses, vetting, minimum scope, and result evaluation.
Keep ML-specific experiment guidance in ml-experimentation, the reference for ML environments, reproducibility, tracking, evaluation, abstractions, and production readiness.
Keep notebook and dashboard guidance in analysis-authoring, the reference for EDA notebook and analytical dashboard composition, visualization selection, and dashboard validation.
Keep AI-system evaluation guidance in evaluation-design, the reference for evaluation dataset design, difficulty balance, metric selection, and tooling fit.
Treat this package as internal foundation knowledge, not a user-selectable workflow.
Preserve the skill authority boundaries and cross-cutting concerns defined in the job registry.

Stop rules

Stop before a state mutation when the session-state protocol has not been loaded or project identity is uncertain.
Stop before a job switch when the user has not confirmed the transition.
Stop and report the missing owner when requested behavior belongs to neither this foundation nor a registered job capability.

Attribution

This package is repository-original orchestration guidance licensed under CC BY 4.0. It does not reproduce or summarize an external standard.

Files

ATTRIBUTION.mdAdded by AutoGineer1021 B
LICENSEAdded by AutoGineer16.6 KiB
SKILL.md3.9 KiB
references/flow-state.md5.3 KiB
references/job-registry.md17.8 KiB
references/lifecycle-classes.md4.0 KiB
references/session-state.md6.9 KiB
references/transition-protocol.md3.0 KiB

references/transition-protocol.md

title: Data Science and Engineering Transition Protocol description: User-confirmed job transition, outgoing disposition, state mutation, and routing rules for Data Science and Engineering Coach sessions

Data Science and Engineering Transition Protocol

Purpose

Use this protocol whenever an explicit request, topic shift, or job completion suggests a different job. A transition changes the foreground job, not the project or coaching-session identity.

Transition sequence

Detect the signal and identify the current and proposed jobs from the job registry.
State the applicable transition rule: the current job's lifecycle class determines what must happen before leaving it.
Name the source job, destination job, expected carryover, and proposed outgoing disposition.
Ask for explicit confirmation. Continue the current job if the user declines or has not answered.
Resolve the outgoing job by class:
- episodic: finish Confirm and use completed, or abandon without a durable partial output and use discarded-cleanly.
- bounded: persist phase and gates and use paused, unless its terminal recommendation was reached and persisted as complete.
- continuous: pass the durable-write gate for pending enrichment, persist it, and use flushed. If the gate blocks, keep the enrichment pending and report that the transition cannot complete yet.
Append one job_log entry containing timestamp, from_job, to_job, rationale, source class, outgoing disposition, and carryover artifacts.
Update current using the session-state mutation rules.
Load the destination's primary skill or confirmed specialist route.
Announce the switch, what carried over, and the destination's next coaching step before asking a destination-specific question.

Outgoing dispositions

Source class	Allowed disposition
`episodic`	`completed`, `discarded-cleanly`
`bounded`	`paused`, `complete`
`continuous`	`flushed`

Do not use a generic switched value. The class-specific disposition is the evidence that resumable or terminal work was handled correctly.

Pause, detour, and return

For a bounded-job detour, persist the source phase and gates before starting the episodic destination. On destination completion, offer to resume the paused job at its recorded phase. Do not resume automatically. If the user chooses another job, record another confirmed transition.

Completion transitions

Completion is not permission to auto-advance. Persist completion first, then offer options such as closing, resuming paused work, enriching continuous work, or selecting another job. A later resume announces completed work but does not re-enter it.

Provenance

This transition protocol is repository-original guidance licensed under CC BY 4.0. It does not reproduce or summarize an external standard.

Requirements and compatibility

Written for: GitHub Copilot, Claude Code

Works with: Claude Code, Claude apps, OpenAI Codex, GitHub Copilot, Cursor, Gemini CLI, Kiro, OpenCode, Goose, Amp, Roo Code, JetBrains Junie, Factory Droid, Devin (ex-Windsurf), Mistral Vibe, OpenHands, Letta

“Works with” means the tool reads the open Agent Skills format; a skill written for one tool may use features others lack.

Needs: Claude Code frontmatter

Security

No issues found by our static scan· No issues found

A static scan is not a guarantee. Skills are third-party content: review a skill before you let an agent use it.

Handoff
Matt Pocock· Community
Generates a concise handoff document to facilitate seamless transitions between AI agents.
Agent toolingClaude CodeClaude appsOpenAI Codex+14 more
MITGitHub stars of the source repository317.5k
Chief Of Staff
Matt Pocock· Community
A strategic orchestration agent that manages long-term goals by coordinating subagents and maintaining a persistent knowledge wiki.
Agent toolingClaude CodeClaude appsOpenAI Codex+14 more
MITGitHub stars of the source repository317.5k
Ask Matt
Matt Pocock· Community
A central routing skill that helps you identify the correct agent skill or workflow for your current situation.
Agent toolingClaude CodeClaude appsOpenAI Codex+14 more
MITGitHub stars of the source repository317.5k
Claude Handoff
Matt Pocock· Community
Seamlessly transfer your current conversation context to a new background agent.
Agent toolingClaude CodeClaude appsOpenAI Codex+14 more
MITGitHub stars of the source repository317.5k
Executing Plans
Jesse Vincent (obra)· Community
Execute implementation plans inline without subagents, maintaining rigorous progress tracking and verification.
Agent toolingClaude CodeClaude appsOpenAI Codex+14 more
MITLow riskHas scriptsGitHub stars of the source repository314k
Using Superpowers
Jesse Vincent (obra)· Community
A mandatory protocol for AI agents to identify and invoke relevant skills before performing any task or responding to user input.
Agent toolingClaude CodeClaude appsOpenAI Codex+14 more
MITGitHub stars of the source repository314k

Download ZIP

The ZIP holds the skill’s files, its licence files and an ATTRIBUTION.md. Download data-science-engineering-foundation.zip