AgentKit
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| POSTGRES_URL | No | PostgreSQL connection string for KPI metrics and declarative tool pack persistence | |
| ANTHROPIC_API_KEY | No | Anthropic API key for multi-provider routing and advanced intelligence analysis | |
| AGENTKIT_ALLOW_WRITES | No | Set to true to enable write and destructive tool pack operations (default: false) | |
| AGENTKIT_APPROVAL_TOKEN | No | Human-held authorization token required for destructive operations such as retract_annotation |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
| logging | {} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| query_kpisB | Retrieve historical and real-time Key Performance Indicators (KPIs) from the governed Postgres store. Fetches structured time-series metrics across enterprise business domains including Finance, People (HR), Operations, Customer (Sales/Support), Engineering, Growth, Logistics, ESG, IT, and Security. |
| get_company_healthB | Compute composite organizational health scores and diagnostic health index breakdown. Synthesizes multi-factor business performance indicators into a holistic 0-100 health score with categorized qualitative interpretations ('Optimal', 'Healthy', 'Warning', 'Critical') and sub-component scores across growth rate, profit margins, cash runway, and operational efficiency. |
| detect_kpi_anomaliesA | Detect statistical anomalies, sudden spikes, and unexpected deviations in KPI time-series history. Analyzes metric historical distributions within a specified business domain using rigorous parametric (Z-Score) or non-parametric (IQR - Interquartile Range) anomaly detection algorithms to identify operational risks, unexpected budget variances, or sudden performance degradation. |
| forecast_metricA | Generate forward-looking time-series forecasts with Monte Carlo confidence intervals for a named KPI. Applies trend analysis and linear projection mathematical modeling over historical monthly metric data to predict future trajectory across specified forward horizon periods, providing upper and lower bound confidence intervals. |
| list_available_metricsA | Discover all supported KPI metrics, business domains, and available historical time periods. Essential discovery tool for agents to inspect the metadata schema before formulating specific queries, anomaly detections, or predictive forecasts. |
| get_executive_summaryA | Generate a comprehensive one-shot executive briefing synthesizing health, key metrics, and anomalies. Executes a high-level cross-domain synthesis combining composite organizational health diagnostics, primary KPI indicators, and active statistical anomalies across Finance, Operations, and Growth into an actionable briefing for leadership. Returns: A dictionary containing: - 'summary': High-level descriptive briefing title. - 'health_score': Overall company health index (0.0 to 100.0). - 'interpretation': Overall company health classification string ('Optimal', 'Healthy', 'Warning', 'Critical'). - 'components': Breakdown of health scores (growth, margin, cash_score, efficiency). - 'key_metrics': Top 5 representative business KPI metric records. - 'anomalies': Top 5 active statistical anomalies requiring immediate executive review. |
| list_annotationsA | Read historical metric annotations and operator notes, ordered from most recent to oldest. Allows filtering by a specific KPI metric and optionally including retracted entries. |
| annotate_metricB | Attach a durable operational finding, investigation note, or root-cause explanation to a specific KPI metric and recording period. Requires write permissions enabled. Effect: WRITE — mutates data. Pass dry_run=true to preview without committing. |
| retract_annotationA | Soft-delete an erroneous or outdated annotation (stamps retracted_at). Classed as a destructive action requiring a valid operator approval token. Effect: DESTRUCTIVE — mutates data. Pass dry_run=true to preview without committing. Requires a human-supplied approval_token. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| monthly_executive_briefing | Reusable prompt template for a monthly cross-domain executive briefing. |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| kpi_finance_latest | Latest Finance KPI snapshot. |
| kpi_people_latest | Latest People KPI snapshot. |
| kpi_operations_latest | Latest Operations KPI snapshot. |
| kpi_customer_latest | Latest Customer KPI snapshot. |
| kpi_engineering_latest | Latest Engineering KPI snapshot. |
| kpi_growth_latest | Latest Growth KPI snapshot. |
| kpi_logistics_latest | Latest Logistics KPI snapshot. |
| kpi_esg_latest | Latest ESG KPI snapshot. |
| kpi_it_latest | Latest IT KPI snapshot. |
| kpi_security_latest | Latest Security KPI snapshot. |
TDQS
Scored across 9 tools
Each tool has a clearly distinct purpose: retraction vs. annotation vs. querying KPIs vs. health scoring vs. anomaly detection vs. forecasting vs. listing metrics vs. executive summary. No two tools overlap in function.
Consistent verb_noun convention (retract_annotation, query_kpis, get_company_health, detect_kpi_anomalies, forecast_metric, list_available_metrics, get_executive_summary, list_annotations, annotate_metric). The verb 'get' vs 'list' vs 'annotate' vs 'retract' is used appropriately based on action type, though 'get_executive_summary' could arguably be 'generate' but it's a read operation.
9 tools is well within the ideal 3-15 range. The set covers discovery, analysis, forecasting, annotation management, and executive summary without redundancy or bloat.
The tool surface covers the full lifecycle for KPI analysis (discovery, query, anomaly detection, forecasting, health scoring, annotations with retraction). However, there is no direct tool for creating or updating KPI metrics themselves (e.g., defining new metrics or modifying existing ones), which may be needed for a governed store. Also, no tool for listing all domains or periods beyond the discovery tool, but that is covered by list_available_metrics.