# Esg Screening Pipeline

> Gather publicly disclosed ESG (environmental, social, governance) signals for a supplied company universe and structure them into a normalized, validated dataset. Use when a researcher needs raw ESG disclosure data assembled for downstream screening; this workflow does not compute a proprietary ESG score or rating.

- Skill: `knuckles-team/esg-screening-pipeline` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add knuckles-team/esg-screening-pipeline`
- Raw SKILL.md: https://api.skillmd.com/api/skills/knuckles-team/esg-screening-pipeline/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: DevOps & Infra
- License: MIT
- Author: Knuckles-Team (https://skillmd.com/u/knuckles-team)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/knuckles-team/esg-screening-pipeline

---


# ESG Screening Pipeline Workflow

Compose the named atomic skills without adding scoring logic here.

## Inputs

Provide the company universe and the ESG disclosure categories of interest.

## Steps

### Step 0: web-search [skill: web-search]

Invoke `$web-search` with the workflow inputs to locate publicly
disclosed ESG reports and filings for the universe.

Expected: `esg_source_packet`

### Step 1: quant-data-ingest [skill: quant-data-ingest] [depends_on: Step 0]

Invoke `$quant-data-ingest` with `esg_source_packet` to normalize the
disclosed data into the Timeseries Memory backend.

Expected: `normalized_esg_dataset`

### Step 2: data-quality-auditor [skill: data-quality-auditor] [depends_on: Step 1]

Invoke `$data-quality-auditor` with `normalized_esg_dataset` and the
declared rule set.

Expected: `quality_report`

## Output

Return `normalized_esg_dataset` and `quality_report`. Does not compute or assert
a proprietary ESG score.

## Execution

- **Run first:** Step 0 — `$web-search`.
- **After level 0:** Step 1 — `$quant-data-ingest`.
- **After level 1:** Step 2 — `$data-quality-auditor`.

**Execution:** If graph-os is reachable, offload the whole DAG via `graph_orchestrate action=execute_workflow` (or the `kg-delegate` skill) for true parallel/swarm execution. Otherwise execute the steps natively in dependency order: run steps with no unmet `depends_on` in parallel, then their dependents.