Start an auditable Hermes task run. Records objective, plan, task graph, autonomy/risk, and initial metadata. Requires x-agent-key.
| Name | Required | Description | Default |
|---|---|---|---|
| plan | No | ||
| context | No | ||
| metadata | No | ||
| objective | Yes | ||
| task_type | No | ||
| risk_level | No | ||
| task_graph | No | ||
| autonomy_level | No | ||
| sub_agents_used | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries the full burden. It discloses the auth requirement and that the tool records objective, plan, task graph, autonomy/risk, and metadata. However, it does not state side effects, return values, or whether the operation is a write/mutation. The word 'Start' implies a state change, but behavioral details beyond recording are absent.
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 exactly three sentences, each earning its place: first states the purpose, second enumerates recorded fields, third notes the auth requirement. It is front-loaded and free of redundancies or unnecessary details.
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?
With 9 parameters, no output schema, and no annotations, this description is too sparse. It omits key parameters (context, task_type, sub_agents_used) and provides no return value or richer behavioral context. While the basics are covered, an agent would likely need additional information to invoke the tool correctly in all scenarios.
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?
Schema description coverage is 0%, so the description must compensate. It mentions objective, plan, task graph, autonomy/risk, and metadata, covering several schema properties. However, context, task_type, and sub_agents_used are not explained, and the description does not clarify the risk_level enum or autonomy_level semantics. Compensation is partial, leaving gaps for several parameters.
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 uses the specific verb 'Start' with the resource 'auditable Hermes task run', clearly distinguishing it from the sibling tool complete_task_run. It also lists the key fields recorded (objective, plan, task graph, autonomy/risk, metadata), adding concreteness. This is a clear, specific purpose statement.
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 does not explicitly state when to use this tool versus alternatives, nor does it mention exclusions. The verb 'Start' implies usage for initiating task runs, and the requirement for 'x-agent-key' is a prerequisite, but no comparison with sibling tools like record_reflection or complete_task_run is provided. Usage guidance is only implied.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Add one secure layer between your agents and this server.
Most task-run and record_* tools are clearly distinct, but query_lessons, query_lessons_full, and query_lessons_premium all describe nearly the same underlying lesson-query action separated mainly by access tier and payment. An agent can easily call the wrong one, especially because 'full' and 'premium' boundaries are unclear.
The tools consistently use snake_case verb_noun names such as start_task_run, complete_task_run, record_reflection, and submit_incident_report. The query_lessons variants use a predictable suffix pattern, and there is no mixing of naming conventions.
Ten tools is a reasonable footprint, but the three query_lessons variants are essentially reimplementations of the same query by payment/access tier. The count is not excessive, though it could be reduced without losing real capability.
The set covers task-run start/completion and lesson querying, but there is no way to read back task runs, reflections, subagent traces, or usage records after submitting them. This makes the task-run side feel mostly write-only and creates a notable gap in a memory-focused server.