data-context-extractor

Name: data-context-extractor
Author: anthropics

2.3kinstalls2.7kforks

How to install data-context-extractor?

anthropics/knowledge-work-plugins/data-context-extractor

$npx -y skills add anthropics/knowledge-work-plugins --skill data-context-extractor

Installs into the current project.

Summary

Generate or improve a company-specific data analysis skill by extracting tribal knowledge from analysts. BOOTSTRAP MODE - Triggers: "Create a data context skill", "Set up data analysis for our warehouse", "Help me create a skill for our database", "Generate a data skill for [comp

Files · 1

View on GitHub

SKILL.md

1	# Data Context Extractor
2
3	A meta-skill that extracts company-specific data knowledge from analysts and generates tailored data analysis skills.
4
5	## How It Works
6
7	This skill has two modes:
8
9	1. Bootstrap Mode: Create a new data analysis skill from scratch
10	2. Iteration Mode: Improve an existing skill by adding domain-specific reference files
11
12	---
13
14	## Bootstrap Mode
15
16	Use when: User wants to create a new data context skill for their warehouse.
17
18	### Phase 1: Database Connection & Discovery
19
20	Step 1: Identify the database type
21
22	Ask: "What data warehouse are you using?"
23
24	Common options:
25	- BigQuery
26	- Snowflake
27	- PostgreSQL/Redshift
28	- Databricks
29
30	Use `~~data warehouse` tools (query and schema) to connect. If unclear, check available MCP tools in the current session.
31
32	Step 2: Explore the schema
33
34	Use `~~data warehouse` schema tools to:
35	1. List available datasets/schemas
36	2. Identify the most important tables (ask user: "Which 3-5 tables do analysts query most often?")
37	3. Pull schema details for those key tables
38
39	Sample exploration queries by dialect:
40	```sql
41	-- BigQuery: List datasets
42	SELECT schema_name FROM INFORMATION_SCHEMA.SCHEMATA
43
44	-- BigQuery: List tables in a dataset
45	SELECT table_name FROM `project.dataset.INFORMATION_SCHEMA.TABLES`
46
47	-- Snowflake: List schemas
48	SHOW SCHEMAS IN DATABASE my_database
49
50	-- Snowflake: List tables
51	SHOW TABLES IN SCHEMA my_schema
52	```
53
54	### Phase 2: Core Questions (Ask These)
55
56	After schema discovery, ask these questions conversationally (not all at once):
57
58	Entity Disambiguation (Critical)
59	> "When people here say 'user' or 'customer', what exactly do they mean? Are there different types?"
60
61	Listen for:
62	- Multiple entity types (user vs account vs organization)
63	- Relationships between them (1:1, 1:many, many:many)
64	- Which ID fields link them together
65
66	Primary Identifiers
67	> "What's the main identifier for a [customer/user/account]? Are there multiple IDs for the same entity?"
68
69	Listen for:
70	- Primary keys vs business keys
71	- UUID vs integer IDs
72	- Legacy ID systems
73
74	Key Metrics
75	> "What are the 2-3 metrics people ask about most? How is each one calculated?"
76
77	Listen for:
78	- Exact formulas (ARR = monthly_revenue × 12)
79	- Which tables/columns feed each metric
80	- Time period conventions (trailing 7 days, calendar month, etc.)
81
82	Data Hygiene
83	> "What should ALWAYS be filtered out of queries? (test data, fraud, internal users, etc.)"
84
85	Listen for:
86	- Standard WHERE clauses to always include
87	- Flag columns that indicate exclusions (is_test, is_internal, is_fraud)
88	- Specific values to exclude (status = 'deleted')
89
90	Common Gotchas
91	> "What mistakes do new analysts typically make with this data?"
92
93	Listen for:
94	- Confusing column names
95	- Timezone issues
96	- NULL handling quirks
97	- Historical vs current state tables
98
99	### Phase 3: Generate the Skill
100
101	Create a skill with this structure:
102
103	```
104	[company]-data-analyst/
105	├── SKILL.md
106	└── references/
107	├── entities.md # Entity definitions and relationships
108	├── metrics.md # KPI calculations
109	├── tables/ # One file per domain
110	│ ├── [domain1].md
111	│ └── [domain2].md
112	└── dashboards.json # Optional: existing dashboards catalog
113	```
114
115	SKILL.md Template: See `references/skill-template.md`
116
117	SQL Dialect Section: See `references/sql-dialects.md` and include the appropriate dialect notes.
118
119	Reference File Template: See `references/domain-template.md`
120
121	### Phase 4: Package and Deliver
122
123	1. Create all files in the skill directory
124	2. Package as a zip file
125	3. Present to user with summary of what was captured
126
127	---
128
129	## Iteration Mode
130
131	Use when: User has an existing skill but needs to add more context.
132
133	### Step 1: Load Existing Skill
134
135	Ask user to upload their existing skill (zip or folder), or locate it if already in the session.
136
137	Read the current SKILL.md and reference files to understand what's already documented.
138
139	### Step 2: Identify the Gap
140
141	Ask: "What domain or topic needs more context? What queries are failing or producing wrong results?"
142
143	Common gaps:
144	- A new data domain (marketing, finance, product, etc.)
145	- Missing metric

People who liked this also viewed

6 picks

Type

data-context-extractor

2.3kinstalls2.7kforks

View on GitHub

How to install data-context-extractor?

anthropics/knowledge-work-plugins/data-context-extractor

$npx -y skills add anthropics/knowledge-work-plugins --skill data-context-extractor

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Data Context Extractor
2
3	A meta-skill that extracts company-specific data knowledge from analysts and generates tailored data analysis skills.
4
5	## How It Works
6
7	This skill has two modes:
8
9	1. Bootstrap Mode: Create a new data analysis skill from scratch
10	2. Iteration Mode: Improve an existing skill by adding domain-specific reference files
11
12	---
13
14	## Bootstrap Mode
15
16	Use when: User wants to create a new data context skill for their warehouse.
17
18	### Phase 1: Database Connection & Discovery
19
20	Step 1: Identify the database type
21
22	Ask: "What data warehouse are you using?"
23
24	Common options:
25	- BigQuery
26	- Snowflake
27	- PostgreSQL/Redshift
28	- Databricks
29
30	Use `~~data warehouse` tools (query and schema) to connect. If unclear, check available MCP tools in the current session.
31
32	Step 2: Explore the schema
33
34	Use `~~data warehouse` schema tools to:
35	1. List available datasets/schemas
36	2. Identify the most important tables (ask user: "Which 3-5 tables do analysts query most often?")
37	3. Pull schema details for those key tables
38
39	Sample exploration queries by dialect:
40	```sql
41	-- BigQuery: List datasets
42	SELECT schema_name FROM INFORMATION_SCHEMA.SCHEMATA
43
44	-- BigQuery: List tables in a dataset
45	SELECT table_name FROM `project.dataset.INFORMATION_SCHEMA.TABLES`
46
47	-- Snowflake: List schemas
48	SHOW SCHEMAS IN DATABASE my_database
49
50	-- Snowflake: List tables
51	SHOW TABLES IN SCHEMA my_schema
52	```
53
54	### Phase 2: Core Questions (Ask These)
55
56	After schema discovery, ask these questions conversationally (not all at once):
57
58	Entity Disambiguation (Critical)
59	> "When people here say 'user' or 'customer', what exactly do they mean? Are there different types?"
60
61	Listen for:
62	- Multiple entity types (user vs account vs organization)
63	- Relationships between them (1:1, 1:many, many:many)
64	- Which ID fields link them together
65
66	Primary Identifiers
67	> "What's the main identifier for a [customer/user/account]? Are there multiple IDs for the same entity?"
68
69	Listen for:
70	- Primary keys vs business keys
71	- UUID vs integer IDs
72	- Legacy ID systems
73
74	Key Metrics
75	> "What are the 2-3 metrics people ask about most? How is each one calculated?"
76
77	Listen for:
78	- Exact formulas (ARR = monthly_revenue × 12)
79	- Which tables/columns feed each metric
80	- Time period conventions (trailing 7 days, calendar month, etc.)
81
82	Data Hygiene
83	> "What should ALWAYS be filtered out of queries? (test data, fraud, internal users, etc.)"
84
85	Listen for:
86	- Standard WHERE clauses to always include
87	- Flag columns that indicate exclusions (is_test, is_internal, is_fraud)
88	- Specific values to exclude (status = 'deleted')
89
90	Common Gotchas
91	> "What mistakes do new analysts typically make with this data?"
92
93	Listen for:
94	- Confusing column names
95	- Timezone issues
96	- NULL handling quirks
97	- Historical vs current state tables
98
99	### Phase 3: Generate the Skill
100
101	Create a skill with this structure:
102
103	```
104	[company]-data-analyst/
105	├── SKILL.md
106	└── references/
107	├── entities.md # Entity definitions and relationships
108	├── metrics.md # KPI calculations
109	├── tables/ # One file per domain
110	│ ├── [domain1].md
111	│ └── [domain2].md
112	└── dashboards.json # Optional: existing dashboards catalog
113	```
114
115	SKILL.md Template: See `references/skill-template.md`
116
117	SQL Dialect Section: See `references/sql-dialects.md` and include the appropriate dialect notes.
118
119	Reference File Template: See `references/domain-template.md`
120
121	### Phase 4: Package and Deliver
122
123	1. Create all files in the skill directory
124	2. Package as a zip file
125	3. Present to user with summary of what was captured
126
127	---
128
129	## Iteration Mode
130
131	Use when: User has an existing skill but needs to add more context.
132
133	### Step 1: Load Existing Skill
134
135	Ask user to upload their existing skill (zip or folder), or locate it if already in the session.
136
137	Read the current SKILL.md and reference files to understand what's already documented.
138
139	### Step 2: Identify the Gap
140
141	Ask: "What domain or topic needs more context? What queries are failing or producing wrong results?"
142
143	Common gaps:
144	- A new data domain (marketing, finance, product, etc.)
145	- Missing metric

People who liked this also viewed

6 picks

Type