Three tools with basic schemas and descriptions, but significant gaps in parameter documentation and output schema clarity. example_tool and search_logs have reasonable descriptions (50-100 chars), but get_commits lacks context on when to use it. All tools have input schemas with type info, but parameter descriptions are minimal or missing for optional fields. No output schemas documented. No error handling guidance. Tool names follow verb_noun pattern (good), but descriptions don't explain prerequisites, return structure, or when to prefer one tool over similar alternatives.

• Document output schemas for all three tools. Example: search_logs should return {logs: [{timestamp, message, level, service}], total_count, next_cursor}. This enables LLMs to chain tools and extract relevant fields.
• Expand get_commits description to 100-150 chars: 'Fetch recent commits from a GitHub repository. Returns commit SHA, author, message, and timestamp. Use this to review recent changes or find when a bug was introduced.'
• Add descriptions to all optional parameters in search_logs. Example for start_time: 'ISO8601 timestamp (e.g., 2024-01-15T10:30:00Z). Logs before this time are excluded. Defaults to 24 hours ago.'
• Add error handling guidance to tool descriptions. Example: 'Returns 404 if repository not found. Try search_repositories() to find the correct owner/repo. Returns 401 if GitHub token is invalid or expired.'
• Document the filters object structure in search_logs. Example: 'Object with Datadog filter keys (e.g., {"service": "api", "env": "prod"}). See Datadog docs for valid keys. Filters are AND-combined with the query parameter.'