Per-repo learning layer for AI coding model routers: learns which task classes can safely run on a cheaper tier. Native for Claude Code, Codex, Antigravity · MCP server · zero deps.
This repository also configures its own agents. See what tierdecay tells them →
Agent Claude Code
TierDecay T1 standard implementation for well-specified tasks: features from a clear spec, unit tests, docs, renames, config, mechanical multi-file edits.
Agent Claude Code
TierDecay T2 complex implementation: multi-file refactors, concurrency, migrations, performance, subtle correctness work.
Agent Claude Code
TierDecay T3 deepest reasoning, read-only: REVIEW of critical diffs (security, auth, money, migrations, public contracts) or SOLVE of novel problems with a complete dossier.
Agent Claude Code
TierDecay T0 read-only reconnaissance: map files, symbols, conventions, and risks before planning. Never modifies anything.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: