StreamKit MCP has 5 tools with basic naming and descriptions, but critical gaps in schema visibility, parameter documentation, and error handling. Tool names follow verb_noun convention (get_*, list_) which is good. However, the source code provided is incomplete, actual tool implementations in streamkit_mcp/tools/*.py are truncated or missing. Only media_tools.py shows a partial schema for get_media_links with 6 parameters (asset_id, width, height, format, quality, crop), all with descriptions. The other 4 tools' input schemas are not visible in the provided code. Tool descriptions are present but brief (34 - 194 chars, within baseline range). No output schemas are documented. Error handling is not evident. Parameter validation rules, enums, and constraints are absent. The code snippet cuts off mid-implementation ('build_media_lin'), preventing full assessment of actual tool behavior.

• Complete and expose input schemas for all 5 tools in the source code. Each schema must include parameter types (string, integer, boolean, enum), descriptions (50 - 150 chars), and constraints (min/max, regex, enum values).
• Document output schemas for all tools. Specify field names, types, and what data each field contains. Example: get_media_links returns {player_url: string, manifest_url: string, thumbnail_url: string, transform_url: string}.
• Add error handling guidance to tool descriptions. State what errors are possible, whether they are retryable, and what the LLM should do next. Example: 'If asset_id not found, try list_recent_assets() to discover valid IDs.'
• Implement pagination for list_recent_assets and list_asset_jobs. Add offset/cursor and total_count to responses. Cap results at 20 - 50 items by default to prevent context window exhaustion.
• Validate all parameters early and return actionable error messages. Example: 'Invalid format: got "png", must be one of: webp, avif, jpeg' instead of a 500 error.