cos_supervise

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

With no annotations provided, the description carries the full burden of behavioral disclosure. It explicitly reveals the critical behavioral trait that the tool never spawns agents, only advises the main agent, which mitigates a key risk. However, it does not detail potential statefulness, side effects on session state, or authentication requirements, leaving some behavioral aspects undisclosed.

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?

Two sentences with no filler. It front-loads the core purpose, then provides a usage sequence and a key constraint. Every sentence adds value and is directly relevant to invoking the tool correctly.

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?

The tool is a complex supervisor with 9 parameters and no annotations, yet the description only covers the high-level loop and one behavioral constraint. It leaves the agent without guidance on how to set the many parameters and what the output schema contains (though an output schema exists). Given the complexity, the description is incomplete for correct invocation of all parameters.

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 has 9 parameters with 0% description coverage, and the description adds no information about what parameters like phase, pending, intensity, dispatched, situation_id, or backtrack_count mean. The only hint is that the tool is called after recording output, but this does not explain the required session_id, task_marker, or persona_id. The description completely fails to compensate for the undocumented schema.

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 function with a specific verb ('Return the next action') and enumerates the possible outcomes (dispatch, backtrack, signal done). It also distinguishes itself from sibling dispatch tools by explicitly stating it never spawns agents itself, making its role unambiguous.

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?

Explicit guidance is provided: 'Call repeatedly after recording each formula output via cos_supervise_record_output' establishes the intended workflow and sequencing. The statement 'Never spawns agents itself — only tells the main agent what to dispatch' clarifies when not to use this tool for actual dispatch, making the usage context very clear.

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