mcp-hr-knowledge-evals
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| EVAL_MODEL | No | Overrides the default model used by the Claude-as-judge scorer. | claude-3-5-sonnet-20241022 |
| ANTHROPIC_API_KEY | No | API key used to enable the Claude-as-judge scoring for the evaluation harness. Without it, evals use the deterministic scorer. |
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 | {} |
| prompts | {} |
| resources | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| query_hr_policyA | Search the HR policy handbook with keyword/semantic matching. Returns verified policy text with citation ids for grounded answers. |
| get_onboarding_milestonesB | Get role-specific 30/60/90-day onboarding expectations and completion criteria. |
| fetch_manager_coaching_rubricB | Synthesize a structured 1-on-1 agenda, coaching prompts, and feedback guidance for a scenario and performance rating. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| conduct_manager_1on1 | Formats HR policy parameters into a system prompt for conducting a policy-grounded manager 1-on-1. |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| Parental Leave Policy | Parental Leave Policy (HR-PL-001) |
| Paid Time Off (PTO) Policy | Paid Time Off (PTO) Policy (HR-PTO-001) |
| Remote Work & Hybrid Policy | Remote Work & Hybrid Policy (HR-RW-001) |
| Travel & Expense Budget Policy | Travel & Expense Budget Policy (HR-TV-001) |
| Onboarding Framework (30/60/90) | Onboarding Framework (30/60/90) (HR-ON-001) |
| Manager 1-on-1 Guidelines | Manager 1-on-1 Guidelines (HR-MGR-001) |
TDQS
Scored across 3 tools
Each tool targets a distinct output: policy text with citations, onboarding milestones, and a coaching rubric/agenda. There is mild conceptual overlap since onboarding expectations and coaching guidance often live inside HR policy, so an agent could occasionally route an onboarding question to query_hr_policy instead.
All names follow snake_case verb_noun structure (query_, get_, fetch_), which is readable and predictable. Minor deviation: 'fetch_manager_coaching_rubric' actually synthesizes content rather than fetching it, so the verb slightly misrepresents the action.
Three tools is on the lean side but each earns its place in a narrow HR knowledge/evals domain with no redundant operations. Scoping is tight rather than padded, though the surface is thin enough that users may wish for one or two more entry points.
For a read-only HR knowledge server, coverage is reasonable: policy lookup, onboarding lifecycle guidance, and manager coaching scenarios. Minor gaps exist, such as no way to enumerate available policy categories or resolve citation ids as a follow-up, but agents can work around these via query_hr_policy.