luum-home/luum-cognitive-os/agent-run-supervision

Trust tier 3
  • community
Source check pending

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

Use when the user asks how a background agent is doing, whether it is stuck/dead, to keep monitoring it, or to produce a handoff from live git/process/WIP evidence.

Install

npx vskill@latest install luum-home/luum-cognitive-os/agent-run-supervision

SKILL.md

Agent Run Supervision

Use this skill to answer background-agent status questions with evidence instead of vibes. It is multilingual and model-agnostic: Spanish, English, and Portuguese triggers should route here.

Quick commands

bash
scripts/cos-agent-run-status --process-id <id> --json
scripts/cos-agent-watch --process-id <id> --interval 60 --max-cycles 5 --json
scripts/cos-progress-metric --process-id <id> --contract progress.yaml --json
scripts/cos-handoff-if-dead --process-id <id> --json

Status states

  • active-progress: process is alive and dirty WIP changed recently.
  • idle-but-safe: process is alive but no fresh WIP signal is present.
  • probably-stuck: same status repeated past the no-progress threshold.
  • dead-with-wip: process is gone while dirty WIP remains.
  • ready-for-handoff: no live process and no dirty WIP blocker.

Multilingual conversational mapping

User phraseAction
como venimos, cómo venimos, status del agenteRun cos-agent-run-status --language es.
seguí monitoreando, dejalo correrRun cos-agent-watch --language es.
murió el agente?, está trabado?Run status; if dead/stuck, run handoff.
how are we doing, agent statusRun cos-agent-run-status --language en.
keep monitoring, let it runRun cos-agent-watch --language en.
como estamos, agente travado, deixa rodarRun with --language pt.

Progress metric contract

yaml
progress:
metric: hir_residual_diff
command: cargo run -- check-hir-parity --json
improves_when: decreases
stuck_after: 3

The metric command should be read-only or validation-only. If it emits JSON, the metric can be selected by key; otherwise the last numeric value in stdout is used.

Rules

  • Do not say an agent is stuck only because it has not committed yet.
  • Check process liveness, dirty WIP age, branch/ahead/behind, validation receipts,

and progress metric before making a status claim.

  • If the process is dead with WIP, produce a handoff instead of editing blindly.
  • If there is recent dirty WIP and a live process, do not overwrite it.
  • Keep answers evidence-backed and mention uncertainty when process detection is by pattern.

Contextual Trigger

Use when the user asks: como venimos, cómo venimos, status del agente, seguí monitoreando, murió el agente, está trabado, dejalo correr, how are we doing, agent status, keep monitoring, is it stuck, did the agent die, como estamos, or agente travado.

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
0

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/luum-home/luum-cognitive-os/agent-run-supervision/badge)](https://verified-skill.com/skills/luum-home/luum-cognitive-os/agent-run-supervision)

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