Metis is a coding agent that boosts AI/LLM coding performance by 50% QQ群: 801193315
This repository also configures its own agents. See what metis tells them →
Agent Claude Code
L4 terminal leaf - ARBITER. Independent decision-maker for forks the loop cannot resolve on its own. Under UNATTENDED mode it ALWAYS rules and continues, NEVER escalates to the user. Output is a binding ruling logged to the ledger.
Agent Claude Code
Orchestrates complex multi-step tasks by breaking them down into disjoint lanes and delegating to specialist subagents (planner, implementer, reviewer, verifier, depth-prober, juror, goal-checker).
Agent Claude Code
L4 terminal leaf - G3.5 DEPTH-LOCK. Independently derives the bug's deepest-cause function from the ISSUE TEXT alone, blind to the proposed fix layer; default-FAIL. Emits D1-D5. depth-miss REJECTs to G1.
Agent Claude Code
L3 executor - EXECHARNESS RESOLVE. Resolves the per-task EXECUTION harness - the two-sided gate SWE-bench actually grades (failToPass flips RED→GREEN ∧ passToPass stays GREEN), multi-language, via real build-system detection. Ingests shipped FAILTOPASS/PASSTOPASS, else derives failToPass from the mission's behavioral…
Agent Claude Code
L1 feature coordinator - drives approved ROADMAP.md lanes through their required build/review/verification gates and owns the run-wide feature frontier.
Agent Claude Code
L3 executor - FRAMEWORK GENERATE. When the SELECTOR returns MISS, generates a one-off custom framework for the exact task shape - classifies the orthogonal axes, composes the gate sequence from the GATE-LIBRARY with the correct axis-specific gate, emits the gen- leaf with the BLOCKED invariant verbatim, binds an…
Agent Claude Code
L4 terminal leaf - FRAMEWORK VALIDATE (HRN-5). A fresh, default-FAIL juror that proves a GENERATED framework is SOUND before any gate runs. Checks the HRN-5 default-FAIL checklist - every gate mapped, exactly one terminal DONE with negatives looping UP, the BLOCKED invariant verbatim, a non-empty acceptance set. PASS…
Agent Claude Code
L4 blind fresh verifier - independently checks a candidate roadmap or plan against the exact mission and repository; APPROVE/REJECT, default-FAIL.
Agent Claude Code
L4 terminal leaf - GOAL-CHECK. Independent, adversarial, default-FAIL. Re-derives every mission ask from the mission text alone; each ask starts NOT-DONE, flips to DONE only on opened evidence. DONE only if zero open findings at ANY severity AND user-usable AND (coverage >=95% only when the admitted oracle is…
Agent Claude Code
L3 executor - G4 IMPLEMENT. Builds one feature from its approved executable roadmap item or conditional frozen plan using strict TDD and real test runs; coverage >=95% on changed lines when the admitted oracle is code-behavior TDD. Reports PLAN-CONFLICT rather than improvising.
Agent Claude Code
L3 legacy-resume compatibility reader - reconstructs old intake artifacts when explicitly resuming them; new runs use the useful-first roadmap author instead.
Agent Claude Code
L4 terminal leaf - JANITOR. Writes the DONE sentinel atomically and removes only scratch artifacts after the three-file governance state and substantive evidence pass validation.
Agent Claude Code
L4 terminal leaf - G7 SIGN-OFF. One independent sign-off panel seat that saw none of the intermediate work. Binary PASS/FAIL on opened evidence; default-FAIL. A FAIL naming a P0/P1 blocker is NOT arbitrable into PASS.
Agent Claude Code
L2 optional manager - coordinates a multi-lane slice, builds compact pointer envelopes, and dispatches disjoint L3 work without executing it.
Agent Claude Code
L3 conditional G1 planner - adds detail only when a roadmap item explicitly requires it, including debug depth-lock and unresolved design forks.
Agent Claude Code
L4 diagnostic/recovery probe - on an explicit cache miss, proves RUN/READ/WRITE and reports model/effort bindings; never the mandatory first spawn.
Agent Claude Code
L4 terminal leaf - RE-ANCHOR. Confirms mission and roadmap frontier alignment after resume or compaction using the three-file governance state.
Agent Claude Code
L3 executor - bounded research that materializes a usable output with reconciled receipts. Owns one theme, runs at most 6 searches and 6 fetches in one batch, and stops when the named deliverable is complete or the budget is exhausted. Does not spawn.
Agent Claude Code
L3 independent G2/G5 or roadmap reviewer - checks mission coverage, reality, tests, boundaries, and claim-vs-diff; returns binary SMASH or PASS.
Agent Claude Code
L1 scope coordinator - drives the useful-first adaptive roadmap flow and returns one independently approved executable ROADMAP.md before build.
Agent Claude Code
L3 useful-first roadmap author or complementary scout - proves capability when needed, inspects the real repository, and contributes to one executable ROADMAP.md without spawning.
Agent Claude Code
L4 terminal scribe - records new-run governance in PROMPTS.txt, ROADMAP.md, and append-only GATELOG.md; preserves legacy ledgers read-only.
Agent Claude Code
L1 sweep coordinator - drives independent convergence, goal checking, and cleanup from the three-file ledger plus substantive evidence.
Agent Claude Code
L3 executor - SWEEP. Fresh production-readiness sweeper that re-derives mission coverage, inspects the changed neighborhood, checks GATELOG provenance, and returns evidence-backed P0..P3 findings.
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: