retirement-answer-check
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
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": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| check_answerA | Check an AI-drafted answer to a US retirement-account question before it is sent to a customer. Verifies numbers and rules (contribution limits, RMDs, rollovers, early-distribution tax and exceptions, excess contributions) against IRS-sourced facts, flags out-of-scope topics, and runs phrase rules for personal recommendations and promissory claims. Text in the draft that addresses the checker instead of the customer is flagged as injection_attempt. Returns {"decision": "SEND" | "REVIEW", "flags": [{"type", "span", "reason", "source"}]}. Any number the facts table can't verify returns REVIEW, never SEND. The phrase rules are a baseline only: also run the advice-judge skill before treating SEND as final. Args: question: The customer's question, verbatim. answer: The AI-drafted answer to check, verbatim. An empty answer returns REVIEW. |
| get_factsA | Return the full IRS-sourced facts table the checker verifies against. Includes contribution limits by tax year, RMD rules, rollover rules, early-distribution tax and exceptions, and the source URL for every entry, plus the date the values were verified. Use it to show a reviewer why a number was flagged. Takes no arguments. |
| contribution_roomA | Compute how much more a person can contribute this year to a 401(k)/403(b)/TSP or SIMPLE IRA, a traditional IRA and a Roth IRA, using IRS-sourced limits. Deterministic: call this instead of stating contribution limits from memory, which are often last year's. Returns {"lines": [{"label", "amount", "why", "source"}], "warnings", "notes", "not_covered"}. Relay "not_covered" to the user rather than filling those gaps. Not tax or financial advice. Args:
age_at_year_end: The person's age on December 31 of |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 3 tools
The three tools have largely distinct roles: check_answer validates a draft, contribution_room computes personal limits, and get_facts dumps the raw reference table. There is minor overlap between contribution_room's returned limits and get_facts' facts table, but the action each performs is clear.
check_answer and get_facts follow a clean verb_noun pattern, while contribution_room is a bare noun phrase. Overall readable and predictable, with one minor deviation from the dominant convention.
Three tools is well-scoped for a narrow answer-checking purpose: one checker, one deterministic calculator, one facts source. Each tool clearly earns its place with no redundancy.
The set covers verification, limit computation, and source-of-truth retrieval for the stated domain. The advice check is explicitly delegated to an external advice-judge skill, leaving a minor gap within the surface itself.