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

Annotations declare destructiveHint=true and idempotentHint=true. The description adds genuinely useful context beyond those: that cancellation preserves changes already in the working directory, which softens the destructive implication, and that it returns the cancellation or existing terminal state, explaining the idempotent 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?

Two sentences, zero waste, and the core action-plus-outcome is front-loaded before the side-effect note. Every sentence earns its place.

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 no output schema, the description responsibly covers the return value ('cancellation or existing terminal state'). For a one-parameter, low-complexity tool this is nearly complete, though it could note whether calling on an already-terminal job errors or silently succeeds.

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 100% and the single job_id parameter is fully documented as 'Job identifier returned by devin_run.' The description adds no syntax or format detail beyond the schema, so the baseline of 3 applies.

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?

States a specific verb+resource ('Stop an existing Devin job') and clarifies the outcome. It is clearly distinguishable from siblings devin_run and devin_wait, which start and poll a job respectively.

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?

Usage is implied by 'Stop an existing Devin job', which tells the agent this is for terminating a running job. However, there is no explicit when/when-not guidance or routing to an alternative, and no prerequisites are named.

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

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

Annotations declare destructiveHint=true and openWorldHint=true, and the description adds critical context: the worker may edit files and run commands (explaining the destructive nature), one external job runs at a time per bridge instance (a concurrency constraint), and it returns a job_id promptly (async behavior). This meaningfully exceeds what annotations provide.

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?

Three tight sentences, front-loaded with the core action and immediately stating the return value and the follow-up tool. No filler.

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?

Covers the async start/poll model, the destructive capability, the concurrency limit, and the result-retrieval path, complementing annotations and the fully-documented schema. No output schema exists, but the description explains the job_id return so nothing essential is missing.

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 coverage is 100%, with each parameter already documented in the schema, including defaults, bounds, and a note that scope does not enforce isolation. The description adds no parameter-level detail beyond the schema, so baseline 3 is appropriate.

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?

States a specific verb and resource ('Start a delegated engineering task'), and explicitly names the sibling devin_wait as the way to collect results, distinguishing it from the polling tool. The async nature and the job_id return are clear.

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?

Clear routing to devin_wait for results and an implied fire-and-forget usage model. However, it doesn't state when NOT to use this tool (e.g., vs. doing the work directly) or describe the role of devin_cancel, leaving some alternative-selection gaps.

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

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

Annotations already establish readOnly, idempotent, non-destructive behavior, so the bar is lower. The description adds blocking semantics (waits up to wait_seconds for completion) and the note that a finished job still needs review, which is genuinely useful context beyond the annotations.

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?

Three short sentences, front-loaded with the core action and no redundant filler. Slightly more compressed would still convey the same, but every sentence carries information.

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 no output schema, the description usefully characterizes the return as a current or terminal result and explains the wait bound. Safety is fully covered by annotations, leaving little an agent needs that is missing.

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 coverage is 100%, so both parameters are already documented by the schema descriptions. The description reinforces wait_seconds as a completion-wait bound but adds no format or edge-case detail beyond the schema, so baseline 3 applies.

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?

States a specific verb (read/wait) and resource (an existing Devin job), with 'existing' implicitly distinguishing it from the sibling devin_run that creates jobs. It does not name devin_run or devin_cancel explicitly, so the differentiation is inferred rather than stated.

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?

Gives practical usage context: reuse the same job_id while running, and review the work even after a completed process. It does not explicitly contrast when to call this versus devin_cancel or devin_run, but the running-job context is clear enough to guide selection.

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