System 1 MCP Server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| TYPESAFE_API_KEY | No | Your TypeSafe API key. If not provided, the server will attempt to resolve it from ~/.system1/config.json, ~/.fastpath/config.json, or a .env file in the working directory. |
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 |
|---|---|
| fast_guardA | Check if a command or action is safe before running it. Returns risk probabilities and a recommended action (pass/review/block). Call this BEFORE executing shell commands, file modifications, or API calls. NOTE: This is advisory -- it returns a recommendation, not enforcement. Args: command: The shell command or action to evaluate. goal: The user's stated task objective / context. workspace: Optional current working directory or project context. block_threshold: Risk probability threshold above which to recommend 'block' (default 0.80). review_threshold: Risk probability threshold above which to recommend 'review' (default 0.40). Returns: JSON object with 'action' (pass/review/block), 'is_destructive', 'is_dangerous', 'is_out_of_scope', and 'blast_radius'. |
| fast_judgeA | Instantly select the best option from a set of candidates. Returns the chosen option with full probability distribution and confidence. Call this when you need to pick one item from a known set without slow deliberation. Args: question: What to decide (e.g. 'Which file contains the database connection settings?'). options: Dict mapping candidate keys to descriptions (at least 2 options). context: Optional context or state relevant to the decision. confidence_floor: Confidence floor (default 0.60) below which 'is_confident' is false. Returns: JSON object with 'choice', 'confidence', 'probabilities', and 'is_confident'. |
| fast_verifyA | Check whether a condition is true or a task goal has been met. Returns the probability (0-1) that the statement is true given the evidence. Call this to verify goal completion, test outputs, status checks, and loop detection. Args: statement: The yes/no claim to verify against evidence. evidence: The output, log, or evidence to evaluate. yes_means: Optional clarification of what 'yes' / true means. no_means: Optional clarification of what 'no' / false means. Returns: JSON object with 'probability', 'is_true', and 'assessment'. |
| fast_scoreA | Rate content along a defined scale with ordered levels. Returns a score, probability distribution across levels, and confidence. Call this for severity, quality, relevance, complexity, or any custom rating. Args: question: What dimension to rate (e.g. 'How severe is this bug report?'). levels: Ordered list of level descriptions from lowest to highest (at least 2). content: The text or object to evaluate. confidence_floor: Confidence threshold (default 0.60). Returns: JSON object with 'score', 'confidence', 'legend', 'probabilities', and 'is_confident'. |
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 4 tools
Each tool has a clearly distinct purpose: guard for safety evaluation, judge for selecting from options, verify for truth checking, and score for rating on a scale. No ambiguity or overlap in their intended use cases.
All tools follow the same 'fast_' prefix plus a descriptive verb (guard, judge, verify, score), forming a consistent and predictable naming pattern throughout the set.
Four tools is well-scoped for a focused server providing fast decision-making helpers. Each tool earns its place, and the count is not excessive or too thin for the declared purpose.
The set covers common fast-decision operations: safety checking, selection, verification, and rating. While a 'compare' or 'rank' tool could be imagined, the judge and score tools suffice for most scenarios, leaving only minor gaps.