Continuous Agent Loop
- 1.4k installs
- 238k repo stars
- Updated August 5, 2026
- affaan-m/ecc
This is a copy of continuous-agent-loop by affaan-m - installs and ranking accrue to the original listing.
continuous-agent-loop is an ECC agent-orchestration skill that runs self-correcting autonomous coding loops with quality gates, evals, and recovery controls for developers who need reliable long-running agent sessions.
About
continuous-agent-loop is an ECC v1.8+ standard loop skill that replaces autonomous-loops while preserving compatibility for one release cycle. The skill routes work through a decision flow—continuous-pr for strict CI/PR control, rfc-dag for RFC decomposition, infinite for parallel exploration, or sequential as the default—and documents a production stack pairing RFC decomposition, plankton-code-quality gates, eval-harness evaluation, and nanoclaw-repl session persistence. Developers reach for continuous-agent-loop when agent runs churn without measurable progress, repeat the same root-cause retries, stall merge queues, or drift costs through unbounded escalation. The skill encodes failure modes and recovery patterns so agent loops terminate with evidence instead of infinite retries.
- Decision tree selects among continuous-pr, rfc-dag, infinite, and sequential loop types
- Recommended production stack combines RFC decomposition, quality gates, eval harness, and session persistence
- Prevents loop churn, repeated root-cause retries, merge queue stalls, and cost drift
- Recovery commands include freeze loop, harness audit, scope reduction, and replay with acceptance criteria
- v1.8+ canonical pattern superseding autonomous-loops with backward compatibility
Continuous Agent Loop by the numbers
- 1,389 all-time installs (skills.sh)
- +82 installs in the week ending Aug 5, 2026 (Skillselion tracking)
- Data as of Aug 5, 2026 (Skillselion catalog sync)
npx skills add https://github.com/affaan-m/ecc --skill continuous-agent-loopAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 1.4k |
|---|---|
| repo stars | ★ 238k |
| Last updated | August 5, 2026 |
| Repository | affaan-m/ecc ↗ |
How do you run reliable autonomous agent loops with quality gates?
Run reliable, self-correcting autonomous agent loops with built-in quality gates, evals, and recovery controls.
Who is it for?
Developers running multi-iteration Claude Code or Cursor agent sessions that need CI-aware PR control, RFC decomposition, or parallel exploration with measurable stop conditions.
Skip if: Developers who only need a single one-shot code edit without loop orchestration, eval harnesses, or autonomous retry management.
When should I use this skill?
An agent session needs a long-running autonomous loop with quality gates, eval loops, or recovery after repeated failures.
What you get
Selected loop strategy, quality-gate checkpoints, eval-harness results, and session-persistent agent run logs.
- Loop strategy selection
- Quality-gate checkpoint plan
- Eval and recovery runbook
By the numbers
- Defines 4 loop strategies: sequential, continuous-pr, rfc-dag, and infinite
- Standard loop skill name in ECC v1.8+ replacing autonomous-loops for one release
Files
継続的エージェントループ
これはv1.8+の標準ループスキル名です。1リリースの間、autonomous-loopsとの互換性を保ちながら置き換えます。
ループ選択フロー
Start
|
+-- Need strict CI/PR control? -- yes --> continuous-pr
|
+-- Need RFC decomposition? -- yes --> rfc-dag
|
+-- Need exploratory parallel generation? -- yes --> infinite
|
+-- default --> sequential組み合わせパターン
推奨される本番スタック： 1. RFC分解（ralphinho-rfc-pipeline） 2. 品質ゲート（plankton-code-quality + /quality-gate） 3. 評価ループ（eval-harness） 4. セッション永続化（nanoclaw-repl）
失敗モード
- 測定可能な進捗なしのループチャーン
- 同じ根本原因での繰り返しリトライ
- マージキューの停止
- 無制限のエスカレーションによるコストドリフト
リカバリー
- ループを凍結する
/harness-auditを実行する- スコープを失敗ユニットに縮小する
- 明示的な受け入れ基準でリプレイする
Related skills
Setup Matt Pocock SkillsScaffold the per-repo configuration that Matt Pocock’s engineering agent skills rely on so they understand the issue tracker, triage labels, and domain documentation la495k196k
Lark AppsConnect Claude, Cursor or custom agents directly to Lark (Feishu) for messaging, document automation, approval workflows and enterprise data access.402k
CavemanSlash token usage by roughly 75% while keeping every technical detail intact when working with Claude Code, Cursor or similar agents.391k94.7k
Lark Skill MakerQuickly turn any Lark/Feishu OpenAPI call or multi-step workflow into a reusable agent skill with its own SKILL.md.386k16k
Running Claude Code Via Litellm CopilotRun Claude Code at a fraction of the cost by routing requests through LiteLLM to the GitHub Copilot Chat API.270k72
Codex PetGenerate a complete Codex Pet spritesheet and metadata from one reference image without needing an OpenAI key or Codex Pro.246k10How it compares
Pick continuous-agent-loop over a single-shot coding skill when agent runs need eval loops, PR-aware control, or RFC-driven decomposition instead of one-pass generation.
FAQ
Which loop strategy should continuous-agent-loop pick?
continuous-agent-loop routes to continuous-pr for strict CI/PR control, rfc-dag for RFC decomposition, infinite for parallel exploration, or sequential as the default when no special constraint applies.
What production stack does continuous-agent-loop recommend?
continuous-agent-loop recommends ralphinho-rfc-pipeline for RFC decomposition, plankton-code-quality with /quality-gate, eval-harness for evaluation loops, and nanoclaw-repl for session persistence.