SkillNarrative Identity
Triage Pre-Graph Data
Audit a dataset or access rule for bad edges before an identity-graph build.
About
Overview
Audit a dataset or access rule before it joins an identity-graph build. Enumerates failure modes (hub identifiers, high-degree nodes, behaviorally suspicious values, over-connected identifiers), tests hypotheses in parallel, quantifies damage by rows / edges / entities, proposes minimal filters ranked by severity, and — when issues are found — authors a validated CREATE MATERIALIZED VIEW that produces a graph-ready clean source. Plans, reports, and authors the clean-view NQL; does not execute it.
Skill Support
Specifications
- Plugin
- Narrative Identity
- Runtime
- Prompt + code
- Version
- 0.5.0
- Languages
- NQL
- Required Tools
- narrative_datasets_describe, narrative_access_rules_describe, narrative_dataset_get_column_stats, narrative_nql_validate
- Included Files
- SKILL.md
Arguments
| Argument | Required / Default | Description |
|---|---|---|
--dataset <id> | Optional | Pre-bind the source as a dataset. Skips discovery. Mutually exclusive with --access-rule. |
--access-rule <id> | Optional | Pre-bind the source as an access rule. Skips discovery. The audit runs identically — access rules behave like datasets in NQL. Mutually exclusive with --dataset. |
--entity-type person|household|device|business|account | Optional | Anchors what "plausible" looks like for the entity being identified. Required if not derivable from context. |
--graph-context standalone|combined | Default: combined | Whether this source will be UNIONed with other sources in the downstream graph build (combined, the common case) or used on its own (standalone). Affects filter recommendations in Phase 6. If not passed and not derivable, the skill asks once in Phase 2. |
--identifier-cols <col,col,...> | Optional | Pre-declare which columns hold identifiers. Skips identifier-column discovery. |
--no-parallel | Optional | Execute hypothesis queries serially (debugging or harness-limited). |
--no-schema | Optional | Work from a user-pasted schema only; skip every narrative-mcp call. |
<free-text tail> | Optional | Additional context about the source system. With no arguments, the skill walks the user through interactively. |
Example prompts
| Prompt | When to use it |
|---|---|
Audit dataset 123 before the graph build | Find hub identifiers and over-connected values that would corrupt connected components. |
Find bad edges in this access rule | Quantify damage (rows / edges / entities) from suspicious graph_edge.target_id values. |
Recommend filters for the graph build | Get a severity-ranked filter list plus a clean CREATE MATERIALIZED VIEW to run. |
/triage-pregraph-data --dataset 123 --entity-type person --graph-context standalone | Audit a person source that ships on its own, anchoring plausibility checks to the standalone graph context. |
Skill Dependencies
This skill builds on the skills listed below — each runs as a step in its workflow.
Requires MCP Servers
This skill calls tools from the MCP servers listed below — connect them so your agent has the tools the skill needs.
Installation