Telos exposes 3 workspace tools with complete input schemas and reasonable descriptions (avg 95 chars). All parameters are typed and described. However, output schemas are not documented, error handling lacks recovery guidance, and descriptions could be more LLM-optimized. Tool naming follows verb_noun convention (workspace_list, workspace_read, workspace_write) but lacks clarity on authorization scope and failure modes. No tool annotations (readOnlyHint/destructiveHint) despite clear risk profiles.

• Document output schemas for all three tools. Example: workspace_list returns {files: [{name, type, size?, modified?}], total_count}. workspace_read returns {content, revision}. workspace_write returns {path, revision, created_at}.
• Add error recovery guidance to descriptions. E.g., 'If root_id is invalid, call workspace_list with a known root_id first. If path not found, check spelling or list the directory.'
• Annotate tools with risk hints: add readOnlyHint=true to workspace_list and workspace_read; add destructiveHint=true to workspace_write. This enables agents to reason about safety.
• Clarify revision semantics in workspace_write description: 'expected_revision prevents concurrent edits. If mismatch, the file was modified by another user, read it again and retry.'
• Add pagination to workspace_list: accept optional limit (default 50, max 1000) and offset parameters; return total_count so agents can iterate through large directories.