eval-tracking
Supabase-backed evaluation tracking with runs, cases, and scores tables. Use when storing eval results, building dashboards, or tracking regression over time.
Find an Agent Skill for your task. Read the guide and copy the install command.
Updated 11-05 00:16
17,405 results
Supabase-backed evaluation tracking with runs, cases, and scores tables. Use when storing eval results, building dashboards, or tracking regression over time.
DigitalOcean Managed Database provisioning and connection patterns
Analyzes on-chain token data including whale tracking, DEX volume, wallet PnL, and token unlock schedules. Use when researching token metrics, portfolio…
Audits whether scores, benchmarks, LLM judges, or surveys measure what we think. Use when a benchmark is saturated, a score unreliable, or judge agreement…
Identify SQL Server performance bottlenecks using DMV queries. Covers wait stats, slow queries, active requests, performance counters, and connection analysis.
Generate concise interactive data science chart HTML from real tables or existing statistical and model results, choose charts according to the question and data conditions, support on-demand interaction and conversational chart modification. Used for viewing data and presenting results; use the corresponding workflow for full analysis reports or application development.
Cash balance monitoring — generate a daily cash balance report every day, verify the consistency of data from three parties, and identify cash risks. Applicable situation: Execute at a fixed time every day (e.g., 9 AM), obtain the balances of each account from [CASH]/[BANK]/[ERP], verify the consistency of data from the three parties, identify balance anomalies and cash risks, and output a daily cash report. Core: three-party reconciliation (cash system/online banking/ERP) + balance anomaly identification + alert triggering.
ESG data integration — Integrate ESG data from multiple sources (environmental/social/governance), check data completeness and accuracy, and generate an ESG data quality report. Applicable situation: at the start of an ESG reporting period (annual/quarterly), or when the user requests "integrate ESG data." Core: multi-source data aggregation + completeness check + data quality scoring.
Insight Generation — Extract insights from business/financial data, identify patterns, anomalies, and opportunities, and transform them into actionable business recommendations. Applicable scenarios: Execute when the user requests to "generate insights," "distill insights," or "find opportunities from data." Core: Data pattern recognition + confidence assessment + the transformation chain from insights to recommendations.
KPI Data Tracking and Reporting — Extract actual KPI data from various business systems, verify data accuracy, and generate multi-dimensional KPI achievement tracking reports. Applicable scenarios: monthly/quarterly KPI tracking cycles, or when the user requests to "track KPI progress." Core: Extract target values vs. actual values from BI/ERP/CRM and other systems + trend analysis.
KPI Management Main Process — Integrates KPI target setting, data tracking, alert handling, and root cause analysis. Applicable situations: Executed during monthly/quarterly business analysis meetings or when KPI anomaly alerts are triggered, integrating kpi-target-setting, kpi-data-tracking, kpi-alert-handling, and kpi-root-cause-analysis, …
KPI Root Cause Analysis — Conduct a systematic root cause analysis of KPIs that have not met targets, trace the true causes, distinguish between systemic factors and one-off factors, and provide a basis for improvement. Applicable situations: when a KPI has not met its target and warning handling cannot resolve it, or when the user requests "analyze KPI root cause." Core: factor decomposition + 5Why analysis + attribution classification + improvement recommendations.
KPI target setting and decomposition — Set KPI target values, decompose company-level targets to business lines/departments, and determine measurement criteria and data sources. Applicable situations: annual/quarterly target setting cycles, or when the user requests "set KPIs" or "decompose targets." Core: target value setting logic (historical baseline/industry benchmarking/strategic backcasting) + decomposition to responsibility units.
Liquidity Coverage Ratio (LCR) Assessment — Assess the company's/group's LCR metrics, identify liquidity risks and propose improvement recommendations. Applicable situations: quarterly liquidity assessment, before and after major transactions, or when the user requests "assess LCR". Core: LCR calculation + composition analysis + risk identification + improvement recommendations.
IFRS 9 three-stage classification — For each credit exposure, perform the SICR (significant increase in credit risk) assessment, determine whether it belongs to Stage 1 / Stage 2 / Stage 3, and record the reason for transfer. Applicable situation: Before IFRS 9 ECL measurement, the risk manager performs Stage classification for each exposure, or performs it when reviewing Stage transfers at quarter-end.
Strategic Assumption Analysis — Identify, document, and track key assumptions in strategic planning, and assess the impact of assumption changes on strategy. Applicable situations: when strategic planning is initiated, during quarterly strategic reviews, or when major events affecting assumptions occur. Core: assumption identification + tracking mechanism + impact assessment + strategic adjustment recommendations.
Comprehensive Valuation Analysis — Integrates multiple valuation methods (DCF/Comparable Company Analysis/Transaction Method), determines the valuation range, and forms a final investment recommendation. Applicable scenarios: M&A investment decisions, or when the user requests a "comprehensive valuation." Core: Multi-method valuation + weight allocation + negotiation range recommendations.
Use ArkCLI to create, validate, query, update, delete, version, and download Datasets. Use when the user mentions Dataset, dataset, ds-*, dsv-*, training set/validation set reuse, data format validation, or connecting a Dataset to a fine-tuning task.
PDF parsing, text extraction, and document format conversion
MUST use when changing a data table's schema, writing or running its migrations, or working with its roles and permissions.
MUST use when creating or modifying a data pipeline, a set of scripts marked `pipeline` and wired together by `on` / `materialize` annotations.
Use when the merchant asks which channels, sources, media, campaigns, or devices are driving revenue, customers, or on-hold pipeline. Produces a…
Use when the merchant asks about customer lifetime value, best customers over time, repeat buyers, one-time versus repeat purchasing, cohort retention, reorder…
Use when extracting tabular data from HTML. Covers GFM Markdown tables, the structured tables array (grid cells plus pre-rendered markdown), and HTML line…