Locus provides 6 well-structured tools with explicit schemas and descriptions. All tools have clear verb-noun naming (locate, refs, hover, diagnostics, status, rename) and documented input parameters with types. However, descriptions are terse (10-50 chars), lacking context on WHEN to use each tool or dependencies between them. Output schemas are not documented in the source. Error handling is basic (generic try-catch). The rename tool lacks confirmation/dry-run pattern despite being destructive. No tool annotations (readOnlyHint/destructiveHint) despite clear risk profiles.

• Expand tool descriptions to 100-200 characters. Include WHEN to use (e.g., 'Use locate to find a symbol definition by name. Use refs to find all usages of a symbol at a specific location.'). Add prerequisites (e.g., 'Requires a workspace with LSP servers initialized').
• Document output schemas for each tool. Add a 'Returns' section to each tool definition showing the structure of results (e.g., refs returns [{filePath, line, character, context}]).
• Add dry_run parameter to rename tool. Return a preview of all changes without applying them. Require explicit confirmation (or a separate apply_rename call) to execute the rename.
• Annotate tools with readOnlyHint and destructiveHint. Mark locate, refs, hover, diagnostics, status as readOnly=true. Mark rename as destructive=true. This enables clients to warn users or restrict agent permissions.
• Improve error messages with recovery guidance. E.g., 'Symbol not found in workspace. Try locate() with a partial name, or check that LSP servers are running (call status()).' Categorize errors: retryable (timeout, service unavailable) vs user-fixable (invalid file path, symbol not found) vs fatal (LSP crash).