debug_session

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description must carry the behavioral disclosure. It reveals that subscribing provides 'buffered + live' traces, indicating the tool receives both historical and new events. It also clarifies the unsubscribe action. However, it does not mention side effects such as whether subscription persists across reconnects, the impact on other subscribers, or any permission requirements. It does not contradict anything, but it leaves room for more explicit disclosure of state changes and lifecycle behavior.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is concise (three sentences) and front-loads the purpose. The first sentence contextualizes the debug system ('Arch connects to live sessions...') which is slightly generic but still relevant. The core function, param usage, and a pointer to a sibling are all included with minimal waste. The structure flows logically from context to action to prerequisite lookup.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a simple subscription-control tool with two required params and no output schema, the description is complete. It covers the action enum, the meaning of each action, and how to obtain the required sessionId. It does not describe the output format (the trace events themselves), but that is not expected for a subscription tool and would likely be covered by the event stream rather than a response. The pointer to debug_list_active_sessions closes the loop on parameter acquisition.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The schema already covers both parameters with descriptions, and the description adds value: it explains what 'subscribe' yields ('buffered + live') and points to debug_list_active_sessions for obtaining a valid sessionId. This goes beyond the bare schema by adding contextual meaning that helps the agent correctly construct calls.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool's core function: 'Subscribe to or unsubscribe from an existing session's trace events.' It identifies a specific verb (subscribe/unsubscribe) and resource (session trace events). It also distinguishes itself from debug_list_active_sessions by pointing to it as a source for session IDs, but does not explicitly differentiate from other debug tools like debug_get_current_state or debug_traces, which could be alternatives for inspecting state rather than subscribing.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides guidance on the two actions ('subscribe' to start receiving traces, 'unsubscribe' to stop) and suggests using debug_list_active_sessions to find session IDs. However, it does not explicitly state when this tool should be used instead of alternative debug tools, nor does it mention any prerequisites (e.g., an active session must exist) beyond the sessionId source. The context implies it is for real-time trace subscription, but exclusions are not stated.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.