Scored across 17 tools
Each tool has a clearly distinct purpose: assess_* variants target different comparison contexts (changeset, side-by-side variant, snapshot, worktree), evaluate_* target different scopes (code, file, project), compare_* target string vs file inputs, and the rest (preference walk, depgraph status/generation, docs, inspect, refactor, begin refactor, coverage) are all unique. No two tools appear to do the same thing.
All tools follow the consistent `topos_` prefix with snake_case and a verb-first pattern: assess_*, evaluate_*, compare_*, plus standalone verbs like inspect, refactor, begin, calculate, generate. The naming is uniform in style and clearly groups related operations.
17 tools is slightly above the typical 3-15 sweet spot, but the scope is broad (evaluation, assessment, comparison, graph management, documentation, coverage, refactoring). Each tool serves a distinct workflow step, so the count is justified without feeling bloated.
The tool surface covers the main lifecycle: evaluating code/files/projects, comparing variants, assessing changes in multiple baseline scenarios, generating and checking dependency graphs, inspecting details, calculating coverage, and obtaining documentation. Minor gaps exist (e.g., no snapshot listing or deletion, no explicit commit-level assessment), but agents can complete core workflows without dead ends.