Skip to content

aiopshwang/data-analysis-ml-agent-skills / skills/running-decision-grade-data-science

running-decision-grade-data-science

Verified

Orchestrate an end-to-end data analysis or machine learning project from decision framing through reproducible handoff. Use when a request spans multiple lifecycle stages or an ambiguous modeling request must become a decision-ready result; use narrower audit or experiment-design skills for isolated reviews.

FileKindSize
SKILL.mdskill md4.4 KB
agents/openai.yamlconfig268 B
assets/decision-grade-project-record.mdmarkdown2.2 KB
references/project-gates.mdmarkdown3.4 KB

assets/decision-grade-project-record.md

View on GitHub
# Decision-grade project record

## Decision contract

| Field | Definition | Evidence or owner |
|---|---|---|
| Decision or action |  |  |
| Decision maker or consumer |  |  |
| Entity and row grain |  |  |
| Target or analytic claim |  |  |
| Prediction or observation time |  |  |
| Outcome window and maturity |  |  |
| Population and exclusions |  |  |
| Primary success measure |  |  |
| Error costs and guardrails |  |  |
| Minimum useful improvement |  |  |
| Operational constraints |  |  |

## Assumptions and material decisions

| ID | Assumption or choice | Why it matters | Evidence or approver | Status | Downstream effect |
|---|---|---|---|---|---|
| A-01 |  |  |  | open |  |

## Input and lineage register

| Input or snapshot | Source and version | Extracted at | Grain and keys | Raw preserved at | Derived outputs |
|---|---|---|---|---|---|
|  |  |  |  |  |  |

## Readiness findings

| Finding | Evidence | Affected rows, entities, or period | Consequence | Treatment | Residual risk |
|---|---|---|---|---|---|
|  |  |  |  |  |  |

Readiness status: **unknown / not ready / ready with conditions / ready**

## Evaluation protocol

- Split and deployment rationale:
- Test-set isolation rule:
- Feature cutoff and leakage controls:
- Primary metric and guardrails:
- Baseline:
- Minimum worthwhile improvement:
- Independent challenge:

## Experiment ledger

| Run ID | Hypothesis | Single material change | Data and split version | Result vs baseline | Relevant slice result | Cost | Decision | Failure note |
|---|---|---|---|---|---|---|---|---|
|  |  |  |  |  |  |  |  |  |

## Claim-to-evidence table

| Claim | Status | Direct evidence | Scope and conditions | Counterevidence or uncertainty | Recommended action |
|---|---|---|---|---|---|
|  | not tested |  |  |  |  |

## Reproduction and handoff

- Reproduction status: reproduced / ready for independent reproduction / replayable with constraints / archival evidence only / not assessed
- Repository revision:
- Immutable input identifiers or checksums:
- Environment or lockfile:
- Configuration and random seeds:
- Command from clean state:
- Expected artifacts and checks:
- Known nondeterminism:
- Owner and review date:
- Monitoring, rollback, or revalidation trigger:
rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 194kUpdated Claude CodeCodexVerifiedInstall