inferencegod

inferencegod/autonomy-loop

Plugin Claude Code

Plugin marketplace listing 1 plugin: autonomy-loop.

inferencegod/autonomy-loop

Plugin Claude Code

Claude Code terminals on one repo pass a git baton: composable builder / reviewer / planner / researcher roles with a deterministic floor under every LLM judgment. v0.8.0 adds a rings + provision-or-refuse security posture: a SessionStart preflight prober classifies an assurance tier (ATTESTED -> LIVENESS -> SEPARATED.

inferencegod/autonomy-loop

Command

Lock the autonomy-loop control plane read-only so a bypassed gate cannot disable itself (v0.8 Ring 1).

inferencegod/autonomy-loop

Command

Scaffold + tune the two-terminal autonomy loop into THIS repo (run once, then approve).

inferencegod/autonomy-loop

Command

Upgrade an existing autonomy-loop install to the current schema (safe, idempotent, run after a plugin update).

inferencegod/autonomy-loop

Command

Terminal 1 - Builder/Ideator autonomy-loop tick (run on a /loop interval).

inferencegod/autonomy-loop

Command

Launch an autonomy-loop terminal inside the srt OS sandbox (allow-only writes + an egress allowlist that also bounds the researcher web-fetch).

inferencegod/autonomy-loop

Command

Terminal 3 - Planner autonomy-loop tick (grill-to-goal spec engine; the feeder in front of the self-feeding loop; run on a /loop interval).

inferencegod/autonomy-loop

Command

Terminal 3 - Researcher autonomy-loop tick (4-terminal OPT-IN; fills the idea pool; run on a /loop interval).

inferencegod/autonomy-loop

Command

Terminal 2 - Reviewer/Fixer autonomy-loop tick (effort-scaled critic panel; run on a /loop interval).

inferencegod/autonomy-loop

Command

ORCHESTRATOR. Run the autonomy loop as subagents in ONE terminal (coordination.mode=single-cli; default OFF; fail-closed). Re-runs decideCoordination, refuses if no independent verification remains, then drives a sequential builder/reviewer baton where the deterministic gate BINDS and the LLM critique is ADVISORY.

inferencegod/autonomy-loop

Hook

Runs when a session starts, executing preflight-run.mjs via node. From inferencegod/autonomy-loop.

inferencegod/autonomy-loop

Hook

Runs before the agent uses a tool for Bash, Edit, Write, MultiEdit and NotebookEdit tool calls, executing gate-guard.mjs via node. From inferencegod/autonomy-loop.

inferencegod/autonomy-loop

Skill Claude CodeCodex

The operating rhythm for the two-terminal autonomy loop - the gate, the honesty mandate, the 5-lens review, the research lane, baton discipline, and the accumulated "what almost broke + the rule that caught it" learnings. Read at the start of every loop session.

inferencegod/autonomy-loop

Agent

You are the BUILDER in a single-CLI autonomy loop. You run as a NAMED subagent in your OWN fresh context and your OWN temporary git worktree. You do not see the parent conversation. You hold the baton for exactly one turn, then hand back. You cannot grant your own pass: a separate reviewer subagent re-runs the…

inferencegod/autonomy-loop

Agent

You are the REVIEWER in a single-CLI autonomy loop. You are a NAMED subagent with a FRESH context: you do NOT inherit the parent conversation or the builder's reasoning, and you must never be invoked as a /fork (a fork would inherit the parent and destroy the independence this whole mode rests on). You have NO write…