run_dsh

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

With no annotations provided, the description carries the full burden and does so thoroughly. It discloses that it can REALLY write files, runs bash/pwsh, is non-interactive with no approval stall, is non-blocking, auto-tracks tasks, has no resume despite accepting session_id, and retries on 429/timeout with exponential backoff. This is far beyond minimal behavioral disclosure.

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 dense but every sentence earns its place: purpose, backend rotation, sandbox boundary, async behavior, no-resume caveat, retry policy, and security warning. The core purpose is front-loaded in the first sentence, and the critical security constraint is placed at the end for emphasis.

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 description covers safety, retry behavior, async execution, and no-resume semantics well, but it does not explain how results are returned or how an agent should retrieve the output of the tracked task. With no output schema and no annotations, and 11 parameters, this is a meaningful gap despite the otherwise rich description.

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 55%, and the description adds meaningful semantics beyond the schema: it explains that workdir is the sandbox boundary, clarifies max_retries as 3 total attempts by default and 0 to disable, and notes session_id is accepted but starts a fresh run. It does not explain timeout_sec, retry_base_ms, retry_max_ms, or task_id, so it does not fully compensate for all undocumented 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?

Opens with a specific verb+resource: 'Async call DeepSeek DSH CLI (dsh --profile headless) to run a task.' It clearly distinguishes from siblings by emphasizing provider-rotatable backends, non-blocking execution, and workspace-write sandbox behavior, which sets it apart from run_codex, run_claude, and run_qwen.

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 gives clear operational context: use it for async, non-interactive tasks that need real file writes and bash/pwsh execution inside workdir. It also provides an explicit guardrail: 'Pass a bounded isolated workdir, never the main repo / a sensitive volume.' It stops short of explicitly naming when not to use it versus specific siblings, but the context is strong enough.

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