cleanexpo/pi-dev-ops/agentic-loop

Claude Code and agent skill by cleanexpo/pi-dev-ops

Trust tier 2
  • community
LLM review pending
Source check pending

Checking that the SKILL.md file is still in the repository.

Infinite self-correcting iteration until completion criteria met. Use when a task needs an open-ended, self-correcting loop that keeps iterating until explicit completion criteria are met, such as an overnight or unattended run, a stop guard that blocks the agent from quitting early, a completion sentinel like PROMISE COMPLETE, or Ralph Wiggum style loops with an iteration cap, an append-only progress file, one small feature per iteration, and a green-CI gate before each commit.

Install

npx vskill@latest install cleanexpo/pi-dev-ops/agentic-loop

SKILL.md

The same SKILL.md was published earlier as cleanexpo/pi-dev-ops/agentic-loop.

Agentic Loop

Two-prompt system: task prompt + stop guard. Agent works -> tries to stop -> guard checks criteria -> not met -> continues.

Safety Rails

  • max_iterations: 20
  • max_tokens: 200000
  • max_runtime_minutes: 60
  • Detect oscillation (fix A breaks B) after 3 iterations

Overnight-loop guardrails (Ralph Wiggum, Pocock 2026-07)

  • Iteration cap: max_iterations above is the backstop against never terminating — keep it.
  • Append-only progress file: each iteration appends learnings ("append", never "update") — memory across context resets.
  • One small feature per iteration: never batch ("do as many as you fancy" recreates the context bloat the loop exists to avoid).
  • Green-CI gate: typecheck + tests pass inside the iteration before its commit; broken commits blind the next fresh context.
  • Completion sentinel: prompt outputs an explicit token (e.g. PROMISE COMPLETE); the loop greps it and exits early rather than relying on the stop guard alone.

[[pocock-ralph-wiggum-overnight-loops-2026-07-14-ingest]]

Security

  • Security scanPassed, 100 of 100Pattern checks on the skill's files.
  • Model reviewNot runA model review is scheduled for the next rescan.

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
5

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/cleanexpo/pi-dev-ops/agentic-loop/badge)](https://verified-skill.com/skills/cleanexpo/pi-dev-ops/agentic-loop)

Something wrong with this skill or its page? Tell us.