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

With no annotations, the description carries the behavioral burden and does so well: it discloses that `css` REPLACES the file, that "" deletes it, that edits are applied all-or-none in order to a single copy, that `old` must match exactly once unless replaceAll (with a whitespace-insensitive fallback), and that a stale expectedHash results in nothing being saved. What it omits is the authorization/permission profile and whether a save creates a restorable version (relevant given the sibling rollback-version/submit-version tools).

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?

Front-loaded with the highest-value fact (what this stylesheet is and when it applies) before the mode selection. Information density is very high for three sentences, though the heavy dash/em-dash punctuation and the trailing `Rules:` fragment make it slightly harder to parse than a cleaner two-block structure would.

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?

An output schema exists, so return values need not be explained, and the description instead covers the things the schema cannot: mode selection, conflict handling, atomicity, and routing to the contract doc and to block-level style tools. It is close to complete; the reliance on an off-tool skill document to convey the full 'contract' is the only remaining gap for an agent operating without that doc.

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 the schema already documents all three parameters; the baseline would be 3. The description adds real meaning beyond it — the mutual exclusivity of `css` and `edits`, the preference order between them, and the provenance of expectedHash (from get-global-css) — which raises it above baseline without fully duplicating schema detail.

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 (write/save) and resource (the SITE-WIDE stylesheet, loaded on every page and in the editor), and immediately delimits scope: rules no single block owns. It names the counterpart tools (get-global-css for the hash) and explicitly routes block-level styles elsewhere, so an agent can distinguish it from the many other save-* siblings 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?

Gives an explicit either/or operating mode (`css` = complete file, or `edits` = fragment replacements), states a preference (`prefer edits for changes`), specifies the prerequisite (pass expectedHash from get-global-css), and adds a when-NOT-to-use rule (block styles belong in the block). It even names a prerequisite read (get-skill-doc {id: "global-css-contract"}) and when to do it (before the first call).

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