ghl_clientportal_cli_sandbox_search

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

The description adds genuinely useful behavior: which file extensions are searched, that dependencies and build output are skipped, that matching is literal, and the OAuth scope (clientportal-ai.readonly). However, the opening boilerplate ('side-effecting: opens/edits sandboxes and starts builds', the `ready` status rule, the 409 explanation) describes the whole /ai-builder-api family rather than this read operation, and an agent may wrongly conclude that running a search mutates the sandbox. It is consistent with readOnlyHint=false but muddies rather than clarifies.

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 actual purpose, file scope, and follow-up call are stated in three tight sentences. But the description is front-loaded with an API-family preamble about opening/editing sandboxes and 409 conflicts that does not apply to a search, so several sentences fail the 'earn its place' test and push the real purpose down.

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?

For a two-parameter read tool with a fully described nested schema and no output schema, the description covers scope, exclusions, matching semantics, and the follow-up read tool. The only real gap is return shape: 'the most matching lines to return' hints at line-level results, but it never says whether file paths and line numbers come back, which matters when the agent must then call `sandbox/view`.

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%: `query`, `paths`, `ignoreCase`, and `maxResults` all carry their own descriptions, examples, and bounds. The prose repeats the literal-match semantics of `query` and the 'whole app when absent' default for `paths` but adds no syntax or formatting detail the schema does not already contain, 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?

The core sentence states a specific verb and resource: 'Search the app's code for literal text.' It further scopes the resource (source files of type ts, tsx, js, jsx, vue, json, md, css, excluding dependencies and build output) and explicitly points at the sibling that reads the results (`POST .../sandbox/view`), so the agent can distinguish it from `ghl_clientportal_cli_sandbox_view` without opening a schema.

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?

It gives a clear workflow: search first, then read matches with `sandbox/view`, and it states a real usage constraint ('matched literally, not as a pattern'), which prevents an agent from passing regex. It does not give explicit when-not-to-use guidance, and the sandbox-editing rules in the preamble do not apply to this particular call.

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