Rebranded to ComposeLint AI. Repository has been moved to TechsOneAi/ComposeLintAI.
No user-submitted related servers found.
Scored across 36 tools
The core token-editing tools are largely distinct, but there is a dense, overlapping cluster of approval/rollback tools (approve_file_changes, approve_change, approve_all_files, reject_file_changes, reject_change, reject_all_files, reject_all_files_force, rollback_last, rollback_pattern) whose boundaries are hard to keep straight. Similarly get_ide_diagnostics vs scan_project_diagnostics and list_tokens vs get_all_token_groups overlap. Descriptions help but the set-level confusion is real.
Almost everything follows a snake_case verb_noun pattern (list_tokens, add_token, run_gradle_build, reject_file_changes). Minor inconsistency in verb choice for analogous operations (list vs get for listings, and various approve/reject/rollback verbs), but overall predictable.
36 tools is well above the comfortable range and the server bundles several distinct concerns (token design, file approval, gradle, sync, diagnostics, config) into one surface. While each concern has real functionality, the volume and the granular approval variants feel excessive rather than tightly scoped.
Coverage of the design-token lifecycle is strong: analysis, detection, edit, impact, approve/reject, and rollback are all present, plus gradle verification, sync, and config mapping. The main gap is no dedicated token deletion tool (orphans must be removed manually), and no explicit string-registry creation beyond set_text, but agents can work around these.