Works with the AI Agents you already use
data lineage and provenance mapper
Data pipelines often become black boxes where the origin of a specific metric is lost across multiple transformations.
$2.99
data lineage and provenance mapper
Example session with this skill installed
Trace the 'Monthly_Churn_Rate' metric from the Retention_Report back to its sources. I have provided the SQL transformation for the 'Calculated_Churn' node and the source table schemas.
- Read your context and instructions
- Compiled the data lineage
I have processed the lineage for the Monthly_Churn_Rate metric based on the provided pipeline metadata. The deterministic trace is complete.
Data Lineage and Provenance Report — Analytics Dept, 2024-05-20
1. Scope, access and sources
- Sources: E1 (Customer_Subscription_Logs), E2 (Billing_Records)
- Access Scope: Internal
- Nodes processed: 3 (2 sources, 1 transform)
- Metrics excluded: 0
- Assumptions: user_supplied (transformation logic)
2. Lineage graph
| id | kind | classification | owner | approved | columns |
|---|---|---|---|---|---|
| E1 | source | internal | DataOps | true | sub_id, status, timestamp |
| E2 | source | internal | Finance | true | sub_id, amount, date |
| T1 | transform | internal | Analytics | true | churn_count, total_base |
Edges for Monthly_Churn_Rate:
Monthly_Churn_Rate->T1.churn_count(Direct)T1.churn_count->E1.status(Conditional: status='cancelled')
3. Provenance ledger
| Metric | Report | Target | Status | Evidence grade | Source columns | Longest path |
|---|---|---|---|---|---|---|
| Monthly_Churn_Rate | Retention_Report | Dashboard_V2 | traced | validated | E1.status, E1.sub_id | 2 hops |
4. Untraceable and blocked measures
None.
5. Findings and items for human review
- Finding INF-01: Information gap — organizational validation required for E2 billing alignment.
- Warning: Transformation uses a hardcoded constant for 'Active' status definitions.
6. Impact analysis
- Impact Request: Change to
E1.status - Downstream:
T1.churn_count(Dist: 1),Monthly_Churn_Rate(Dist: 2). - Owners: Analytics Team, Retention Lead.
7. Proposed actions (drafts), assumptions and limitations
- Action: Approve
Monthly_Churn_Ratefor executive review (status: draft). - Observation: Computed result shows 100% dependency on E1 subscription logs.
This analysis supports, and does not replace, professional judgment and management's decision.
Management decision required for final publication.
Next steps
- Review the hardcoded constants flagged in section 5.
- Confirm the
E2billing record owner approves the joined logic. - Validate the
churn_countexpression against the latest business requirements.
Connects securely to your tools. The creator never sees your data.
About this skill
Data pipelines often become black boxes where the origin of a specific metric is lost across multiple transformations. This skill reconstructs column-level lineage and provenance using your pipeline metadata, ensuring every published figure is traceable to an approved source.
What it does
- Lineage Reconstruction traces metrics from reports back through transformations to original source columns.
- Provenance Mapping generates a ledger detailing source approval status, validation grades, and fallback rules.
- Integrity Flagging identifies untraceable measures or metrics drawing from unverified and unapproved sources.
- Impact Analysis predicts downstream effects on reports and metrics when a source column or table is modified.
- Access Verification checks that metrics do not expose sensitive or confidential columns beyond the defined access scope.
How it works
- Metadata Collection involves registering sources with unique identifiers (E-numbers) and defining node transformations.
- Request Formatting builds a
request.jsonfile following a strict schema including nodes, columns, and access scopes. - Deterministic Parsing runs a specialized Python script to build the dependency graph and trace each metric logic.
- Report Generation produces a structured provenance report with findings, severity codes, and items for human review.
Frameworks & tools
The skill utilizes a self-contained Python execution environment to process JSON-based metadata schemas. It operates entirely on supplied metadata without external database connections.
Why this beats prompting it yourself
Manual tracing of nested SQL expressions and cross-table dependencies is prone to hallucination in standard LLMs. This skill uses a deterministic script to ensure lineage is calculated rather than guessed, providing exact evidence grades and source references.
Use cases
- Auditing a financial report to verify that all figures originate from approved ERP tables.
- Performing a change impact assessment before refactoring a core data warehouse table.
- Identifying "shadow data" metrics that rely on unvalidated or estimated inputs.
- Mapping data sensitivity flows to ensure PII is not leaked into public-facing dashboards.
Known limitations
This skill only analyzes provided metadata and does not connect to live databases or file systems. It requires explicit transformation logic or column dependencies to be defined in the input.
Requires manual supply of metadata via JSON. No live system connectivity. Does not verify if the actual data matches the metadata. Interpretation is restricted to the provided access scope.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 2 days ago