continuous-agent-loop

Decision framework for autonomous agent loops with quality gates, evals, and recovery patterns. This is the v1.8+ canonical loop skill name. It supersedes autonomous-loops while keeping compatibility for one release. Recommended production stack: - Four loop modes available: continuous-pr for CI/PR workflows, rfc-dag for RFC decomposition, infinite for exploratory parallel generation, and sequential as default - Recommended production stack combines RFC decomposition, code quality gates, eval harness, and session persistence for robust autonomous execution - Built-in failure mode detection covers loop churn, repeated retries, merge queue stalls, and cost drift with explicit recovery procedures - Recovery workflow includes loop freezing, harness audits, scope reduction, and replay with acceptance criteria to handle degraded states - RFC decomposition ( ralphinho-rfc-pipeline ) - quality gates ( plankton-code-quality + /quality-gate ) - eval loop ( eval-harness ) - session persistence ( nanoclaw-repl ) - loop churn without measurable progress - repeated retries with same root cause - merge queue stalls - cost drift from unbounded escalation - freeze loop - run /harness-audit - reduce scope to failing unit - replay with explicit acceptance criteria