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

Annotations already declare destructiveHint=true and openWorldHint=false, so the agent knows this is an unconfined-but-sandboxed mutation surface. The description adds the useful context that execution is confined to the 'project sandbox', but says nothing about output capture (stdout/stderr/exit codes), timeout behavior, or what state a destructive command can affect.

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 short sentences with zero waste; the core action is front-loaded and the caveat follows immediately.

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?

This is a destructive, four-parameter execution primitive with no output schema, yet the description never explains what a call returns or how failures are surfaced. Given the untyped `command` string and undocumented timeout semantics, an agent is left guessing on important operational details.

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 only 25% — only `project_id` is documented, while `command`, `cwd`, and `timeout_ms` have no schema text. The description contributes nothing about any parameter; it doesn't even state that `cwd` defaults to the project root or how `timeout_ms` interacts with `bash_await`.

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 — 'Run a shell command in the project sandbox' — which an agent can distinguish from file tools or preview tools. It doesn't explicitly differentiate itself from the sibling `bash_await` or `browser_run`, so sibling routing is only partially covered.

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 second sentence gives one useful exclusion: long-lived preview servers belong to the project run config, not a shell. However, there is no guidance on when to use `bash` versus `bash_await` for long-running work, nor any note about preferring `edit_file`/`write_file` for file changes.

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