cwnu-open-source-community/universal-code/loop-monitor

Claude Code and agent skill by cwnu-open-source-community/universal-code

Trust tier 3
  • community
Source check pending

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

Install

npx vskill@latest install cwnu-open-source-community/universal-code/loop-monitor

SKILL.md

Loop monitor workflow

Monitor only an explicit target Run. Default to Root in Plan with read_only execution, a hard deadline, maximum rounds, elapsed budget, retry/backoff policy, stop-on-terminal behavior, and a notification policy. Use a zero model-call budget unless the operator deliberately chooses a bounded positive budget. A periodic schedule must also declare an elapsed interval and IANA timezone; a one-shot schedule has exactly one occurrence.

Create durable state with cyberagent run schedule create <run-id> --at <RFC3339> --deadline <RFC3339> --operation-key <stable-key> and add --every <duration> only for a periodic monitor. Inspect with cyberagent run schedule list <run-id> and cyberagent run schedule show <job-id>. Pause, resume, or cancel with the current revision and a stable operation key. run schedule tick is a single foreground reconciliation step, not an unbounded agent loop or service install.

Each round observes only persisted metadata for the target Run. If the observation digest is unchanged, record an unchanged round without calling a model or tool. Stop when the target is terminal, the deadline or elapsed budget is reached, rounds/model calls are exhausted, authorization becomes stale, retries are exhausted, or the operator cancels.

Use cyberagent doctor snapshot --run <run-id> --json and a bounded cyberagent debug query --run <run-id> --limit 100 --json for diagnosis. Event payloads, prompts, terminal input, command input, secrets, lease-owner values, and fence tokens are never monitoring output.

An approved_repair schedule is exceptional. It requires Code/Deliver plus an exact, unexpired mode and execution-permission authorization bound to this job. It may use only separately offered repair capabilities and ordinary approval checks. Monitoring cannot create, refresh, bypass, or widen that authority, and must fail closed when the binding changes.

Treat this Skill as guidance only. It grants no model, tool, filesystem, process, network, sandbox, persistence, notification, or approval authority and does not enable autostart or a remote control plane.

Security

  • Security scanPassed, 90 of 100Pattern checks on the skill's files.
  • Model reviewPassed

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/cwnu-open-source-community/universal-code/loop-monitor/badge)](https://verified-skill.com/skills/cwnu-open-source-community/universal-code/loop-monitor)

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