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

Annotations only flag it as a non-idempotent, open-world write; the description adds real behavioral context beyond that — duplicate detection returns candidates instead of creating, same-site conflicts block creation without explicit consent, from_preview_id alone is a read-only prefill while submitting with observed_website_link creates transactionally, and the observed URL preserves a concurrency guard. That is meaningful disclosure of gating and side effects rather than restating hints.

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 core action and its precondition are front-loaded, and every sentence carries a distinct rule or fallback branch. The compressed, token-dense phrasing ("same_site_project conflict requires explicit link_existing:true") is efficient but reduces first-read clarity slightly.

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 seven-parameter, no-required-parameter mutation with no output schema, the description covers the ambiguous branches (duplicates, same-site conflicts, preview prefill) and the checkout follow-up. It does not describe what a successful creation returns (e.g. resulting project_code) or any auth prerequisite, which are the remaining gaps.

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 3 is the baseline, but the description genuinely adds cross-parameter workflow meaning: from_preview_id reads prefill and then must be paired with observed_website_link to create, and link_existing is only meaningful in response to a same_site_project conflict. The duplicate_choice and project_code pairing is also explained as a consent-driven interaction rather than as isolated fields.

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 ("Create a project") plus the precondition (after approval of reference label and published app URL), and disambiguates project_name from the home-screen app name. It never names a sibling tool, so routing versus bdk_projects or bdk_setup_preview is left to inference.

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 explicit conditional paths: duplicates return candidates and the agent must ask the customer to choose create_new vs use_existing, a same_site_project conflict requires explicit link_existing true/false, and preview should be offered before checkout. It also names the alternative action (use_existing with project_code) and forbids the tempting shortcut ("never reuse automatically").

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