Agentic Wallet Guardian
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| GUARDIAN_API_KEY | No | API key for request authentication. If not set, authentication is disabled. | |
| GUARDIAN_TX_BUILDER | No | Transaction builder: 'mock' (default) or 'rpc' to build real calldata for transfer/approve intents. | |
| GUARDIAN_STORAGE_BACKEND | No | Storage backend: 'memory' (default, in-memory) or 'sqlite' (persistent). | |
| GUARDIAN_RPC_ESTIMATE_AGE | No | Set to 'true' to enable wallet age estimation via RPC (requires an archive-capable node). Defaults to 'false'. | |
| GUARDIAN_SIMULATION_PROVIDER | No | Simulation provider: 'mock' (default, no network) or 'rpc' (real eth_call dry-run). | |
| GUARDIAN_RATE_LIMIT_PER_MINUTE | No | Maximum number of requests per minute for the API. Leave unset for no rate limit. |
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 |
|---|---|
| evaluate_actionA | Evaluate a proposed blockchain action before executing it. Call this BEFORE submitting any transaction on behalf of a user or agent. Returns a JSON object with a decision ("ALLOW", "WARN", or "BLOCK"), a 0-100 risk score, and a human-readable explanation of which signals drove the decision. Treat "BLOCK" as a hard stop and "WARN" as something that needs explicit user confirmation before proceeding. Args: agent_id: Stable identifier for the calling agent (used for reputation history - reuse the same id across calls). wallet: The wallet address initiating the action. chain: Chain name, e.g. "ethereum", "base", "arbitrum", "polygon". action_type: One of "swap", "transfer", "approve", "contract_call", "bridge". target: Contract or recipient address, if applicable. from_token: Token symbol being sent/sold, if applicable. to_token: Token symbol being received/bought, if applicable. amount: Transaction amount in the relevant unit. |
| get_agent_historyA | Return this agent's past decisions and current reputation score. Useful before evaluate_action if you want to check standing first, or after a WARN/BLOCK to review why. |
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 2 tools
Each tool has a clearly distinct purpose: one evaluates a proposed action, the other retrieves an agent's decision history. No overlap exists between them.
Both tool names follow the consistent verb_noun pattern (evaluate_action, get_agent_history), making the API predictable and easy to navigate.
With only two tools, the server feels thin for a watchdog service. While the focus is narrow, users might expect additional management or reporting tools, making it borderline.
The core lifecycle of evaluating an action and reviewing past decisions is covered. Minor gaps exist, such as no explicit policy management or post-execution reporting, but these are not critical for the stated purpose.