菜单
Query and explore Google BigQuery datasets using the REST API. Use this skill when the user asks to run SQL queries against BigQuery, explore datasets/tables, check table schemas, or estimate query costs. Triggers on BigQuery, BQ, or data warehouse requests.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Full read-write access to Plain customer support platform. Create and manage customers, threads, companies, tenants, labels, tiers, help center articles, and more. Designed for workspace setup and demo preparation.
Read, write, and manage Google Sheets spreadsheets using the REST API. Use this skill when the user asks to read or update spreadsheet data, search for spreadsheets, manage sheet tabs, or format cells. Triggers on Google Sheets, spreadsheet, or sheet-related requests.
Query and browse Intercom conversations, contacts, tickets, and tags via the Intercom REST API. This skill should be used when a user wants to search conversations, look up contacts, find tickets, list tags, or read full conversation threads in their Intercom workspace. Requires an Intercom API access token stored as INTERCOM_API_TOKEN in the environment or vault.
Attio CRM operations for companies, contacts, and notes management. Use this skill when interacting with Attio CRM, searching companies, creating contacts, adding notes, or managing deal records. Triggers on CRM operations, company lookup, contact management, or Attio-related requests.
Access to Plain customer support platform. Read customers, threads, timeline, and help center content. Add notes to threads. Create, update, and publish help center articles.
| name | bigquery |
| description | Query and explore Google BigQuery datasets using the REST API. Use this skill when the user asks to run SQL queries against BigQuery, explore datasets/tables, check table schemas, or estimate query costs. Triggers on BigQuery, BQ, or data warehouse requests. |
Query and explore Google BigQuery directly via the REST API. No SDK or bq CLI needed — the bundled script handles service account authentication and calls the BigQuery v2 REST endpoints using only Python stdlib.
What do you need?
│
├── Explore what data exists
│ └── python scripts/bigquery_api.py list-datasets
│ └── python scripts/bigquery_api.py list-tables <dataset>
│
├── Understand a table's structure
│ └── python scripts/bigquery_api.py get-schema <dataset> <table>
│
├── Run a SQL query
│ └── python scripts/bigquery_api.py query "SELECT ..." --max-results 100
│
└── Estimate query cost before running
└── python scripts/bigquery_api.py query "SELECT ..." --dry-run
# Required — set to the base64-encoded JSON content of a Google Cloud service account key
# Generate with: cat service-account.json | base64 -w 0
GOOGLE_SERVICE_ACCOUNT_KEY='eyJ0eXBlIjoic2VydmljZV9hY2NvdW50IiwicHJvamVjdF9pZCI6Ii4uLiIsInByaXZhdGVfa2V5IjoiLi4uIiwiY2xpZW50X2VtYWlsIjoiLi4uIn0='
The value must be the base64-encoded version of the full service account JSON key file. This avoids shell escaping issues with raw JSON containing special characters.
The service account needs the BigQuery User role (roles/bigquery.user) at minimum. For write access, add BigQuery Data Editor.
Generate a key: Google Cloud Console > IAM & Admin > Service Accounts > Keys > Add Key > JSON. Then base64-encode it before setting the env var.
python scripts/bigquery_api.py list-datasets
python scripts/bigquery_api.py list-tables my_dataset
python scripts/bigquery_api.py get-schema my_dataset my_table
python scripts/bigquery_api.py query "SELECT * FROM my_dataset.my_table LIMIT 10"
python scripts/bigquery_api.py query "SELECT * FROM my_dataset.my_table" --max-results 50
python scripts/bigquery_api.py query "SELECT * FROM my_dataset.big_table" --dry-run
Returns estimated bytes to be processed without actually running the query. Useful before running expensive queries on large tables.
When working with an unfamiliar BigQuery project, follow this sequence to build context:
list-datasets — see what datasets existlist-tables <dataset> — see tables and row countsget-schema <dataset> <table> — inspect column names, types, and descriptionsquery "SELECT * FROM dataset.table LIMIT 5" — sample actual dataFor projects with many tables, use INFORMATION_SCHEMA to get an overview:
python scripts/bigquery_api.py query "SELECT table_name, row_count, size_bytes FROM project.dataset.INFORMATION_SCHEMA.TABLE_STORAGE ORDER BY size_bytes DESC"
LIMIT clause when exploring unfamiliar tables--dry-run before running queries on large tables to check costSELECT specific_columns over SELECT * to reduce bytes scannedWHERE date >= '2026-01-01') to limit scan scopeThe script handles async queries automatically. If BigQuery returns jobComplete: false, the script polls for results with a 2-second interval and a 120-second timeout. No manual intervention needed for queries that take a few seconds to complete.
GOOGLE_SERVICE_ACCOUNT_KEY contains a base64-encoded private key — store it in a vault, never in source controlSet the environment variable to the base64-encoded JSON content of the service account key file (cat key.json | base64 -w 0). On Nairi, store it as a vault secret.
The service account lacks BigQuery permissions. Assign BigQuery User role in Google Cloud IAM.
Check dataset and table names with list-datasets and list-tables. Names are case-sensitive.
Queries over 120 seconds will time out. Add filters or LIMIT to reduce the result set, or break the query into smaller parts.
The script needs either the cryptography Python package or openssl CLI to sign JWTs. Most systems have at least one. Install with pip install cryptography if needed.