This server has 12 tools with significant quality gaps. Tool naming is inconsistent (session_* vs shannon_*), and descriptions vary widely in quality. Most critically, 4 tools (shannon_auth_session, shannon_browser, shannon_crawler, shannon_headers_audit) have complex schemas with multiple conditional parameters that lack clear documentation of dependencies. Error handling is absent, no guidance on retryability, user-fixable vs fatal errors, or recovery paths. Output schemas are not documented. Parameters like 'action' enums are present but lack descriptions of what each action does. Security concerns: shannon_exec and shannon_exploit expose arbitrary command execution with minimal guardrails. The server mixes session management (session_create, session_list, session_switch) with security testing tools, but the relationship between sessions and tool execution is unclear.

• Document shannon_auth_session parameter dependencies explicitly: 'For action=create, provide target and session_type. For action=get/delete/build_headers, provide session_id. For action=list, no additional params required.' Add this to the tool description.
• Add output schema documentation to every tool. Example for session_create: 'Returns {session_id: string, session_name: string, created_at: ISO8601, status: string}'.
• Implement confirmation/dry-run for shannon_exec and shannon_exploit. Add a 'dry_run' boolean parameter (default false) that echoes the command without executing, or require explicit 'confirm: true' parameter for destructive operations.
• Add error handling guidance to tool descriptions. Example: 'Timeouts (>5min) are retryable; permission errors are fatal. If auth fails, verify credentials via shannon_auth_session first.'
• Clarify session context in tool descriptions. Document whether shannon_exec runs in the active session, or if session_id must be passed explicitly. Add session_id as an optional parameter to all shannon_* tools if context-dependent.