No user-submitted related servers found.
Scored across 9 tools
The core tools target distinct resources: project inventory (lister_projets), single-project deep dive (decrire_projet), document content (lire_cerveau), git state (etat_git), content search (chercher_partout), filename search (trouver_fichiers). The only real overlap zone is the overview trio (apercu_ecosysteme, lister_projets, etat_git without a project), but descriptions explicitly delimit them ('plus dense que lister_projets', 'n'inclut volontairement ni l'état git') and even prescribe the ordering, so an agent can pick correctly.
Casing and language are perfectly uniform (all snake_case, French), but the convention is mixed: five tools use a verb-led pattern (lister_projets, decrire_projet, lire_cerveau, chercher_partout, trouver_fichiers) while four are noun phrases (apercu_ecosysteme, etat_git, diagnostic_pilotage, usage_ia_claude). Still highly readable and predictable in tone, but not a single consistent verb_noun scheme.
Nine tools is right in the sweet spot for a machine-wide observability server, and each covers a distinct axis (inventory, detail, docs, git, content search, path search, diagnostics, cost usage). No filler tools and no obviously missing companion tool.
The read-only 'pilotage' domain is well covered end to end: discovery, per-project detail, context docs, git status, two complementary search modes, self-diagnostics, and API cost tracking. The notable gap is generic file reading — trouver_fichiers and chercher_partout can locate a file but there is no tool to read arbitrary file contents (lire_cerveau is limited to markdown context docs), and there is no diff/history inspection beyond etat_git's commit list.