Single tool with minimal schema documentation. Tool name 'compress_prompt' is clear and verb-based, but input schema lacks parameter descriptions entirely. The schema object shows properties but no 'description' fields for 'text' or 'target_ratio' parameters. Tool description is present (58 chars) but generic, does not explain WHEN to use it, WHAT it returns, or any prerequisites. No output schema documented. No error handling guidance. Fails multiple critical checks from pattern:tool and pattern:tool-description.

• Add descriptions to all input parameters in the schema. For 'text': 'The prompt or document to compress (string, any length).' For 'target_ratio': 'Compression ratio as decimal 0.0 - 1.0 (default 0.5). E.g., 0.5 keeps ~50% of tokens, 0.3 keeps ~30%. Lower = more aggressive.'
• Expand tool description to 100 - 150 chars explaining WHEN and WHY: 'Prune low-information tokens (stop words, articles, prepositions) from long prompts to reduce context window usage and API costs while preserving key entities, numbers, and imperatives. Returns original and compressed token counts plus savings.'
• Document the output schema explicitly in the tool description or as a separate schema field. E.g., 'Returns: {compressed_text (string), original_tokens (int), compressed_tokens (int), compression_ratio (float 0 - 1), tokens_saved (int)}'.
• Add error handling with actionable messages. E.g., if text is empty, return 'Error: text cannot be empty.' If target_ratio is outside 0 - 1, return 'Error: target_ratio must be between 0.0 and 1.0 (got X).'
• Add a note in the description about edge cases: 'Texts with ≤5 tokens are returned unchanged. Compression respects original word order; output is a subsequence of the input.'