Four tools with complete input schemas and descriptions, but significant gaps in parameter documentation and output schema clarity. Tool names follow verb_noun convention (foldo_freeze, foldo_replay, foldo_apply_edit, foldo_list_branches). Descriptions are present (40-80 chars) but lack actionable context for LLM selection. Parameters like 'recipe' (array type) and 'target' (object type) lack detailed descriptions explaining structure and constraints. No documented output schemas visible. Error handling and recovery guidance absent. Naming is clear but descriptions fall short of the 50-200 char LLM-optimized baseline.

• Expand tool descriptions to 50-150 chars. Example: 'foldo_freeze: Capture a screenshot of the running app at a specific branch/commit/route and post it to the Foldo canvas. Use this to document UI state changes across commits. Requires valid boardId, branchId, commitSha, and route.'
• Document nested parameter structures. For 'recipe' array, specify: 'Array of UI recipe steps. Each step is an object with type (click|fill|navigate), selector (CSS selector string), and value (for fill steps). Example: [{type: "click", selector: ".submit-btn"}, {type: "fill", selector: "input[name=email]", value: "[email protected]"}]'
• Add output schema documentation to each tool. Example for foldo_freeze: 'Returns {frameId: string, boardId: string, timestamp: ISO8601, imageUrl: string, stateLabel?: string}. frameId can be passed to foldo_apply_edit as reference.'
• Add explicit state-modification warnings to WRITE tools. foldo_freeze and foldo_apply_edit descriptions should state: 'This operation creates a new commit and modifies the repository. It is not idempotent, repeated calls with identical parameters will create duplicate commits.'
• Document error scenarios and recovery paths. Example: 'If commitSha is invalid, returns {error: "commit_not_found", available_commits: [...]}. Try foldo_list_branches to discover valid commits.' Add similar guidance for all tools.