本文へ移動
cccskills
無料GitHub で公開

long-task-watcher

Reliable progress watching for long-running background commands (training runs, live audits, batch jobs, CI) without losing output to stdout buffering. Prefer post-completion cat-and-grep over tail -F.

インストール方法を見る

含まれるファイル(1)

  • SKILL.md1.9 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Long-Task Watcher

Patterns for tracking a long-running background command and surfacing its events without the classic tail -F | grep buffering trap: matching lines can sit in a 4KB stdout block buffer and never emit within the watch window, so a watcher waits forever on output that is already on disk.

Pattern selection

SituationPatternWhy
Task finishes within minutesWait for completion, then cat <logfile> | grep -E "<filter>"No inotify/buffering dependency; no lost lines. Default choice.
Long task, frequent progress linesstdbuf -oL tail -F <logfile> | stdbuf -oL grep --line-buffered "<filter>"Forces line-buffering at every pipe stage (macOS: needs coreutils).
External state (CI, deploy, queue)Polling loop: while true; do <one-shot status call>; sleep 30; doneThe external status API is the truth; tailing local output adds nothing.
Task should stop the watch when it diesAdd kill -0 $PID 2>/dev/null || break to the polling loopBreaks immediately on process exit instead of waiting for a timeout.

Rules

  • Redirect long-run stdout/stderr to a file from the start (... > run.log 2>&1); never rely on scrollback.
  • After the task exits, one cat-and-grep pass over the log is more reliable than any live tail.
  • Watch loops need a hard timeout; a watcher without one outlives dead tasks.
  • Archive logs to a stable location (.audit/ convention) before summarizing, so the evidence survives the session.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Audit agent runtimes and their tests for lifecycle inference, capability overload, context pollution, correlated verification, prompt-only control, false success, dead compatibility paths, test-only seams, and complexity without contribution. Use for whole-repository agent audits, slop or dead-code triage, test contribution reviews, large refactors, tool/context simplification, and cleanup requests that need proof before deleting code or tests.

日本語の概要は準備中です。原文の説明を表示しています。

mangowhoiscloud/geode152026年10月12日 更新

Diagnose agent operational failures through default-policy checks, multi-gap analysis, explicit service ownership, request-context lifetime, and degradation-versus-correctness checks.

日本語の概要は準備中です。原文の説明を表示しています。

mangowhoiscloud/geode152026年10月12日 更新

Align, run, audit, and publish GEODE benchmark comparisons against Agent-World arXiv 2604.18292v1 across MCP-Mark, BFCL V4, and tau2-bench. Use for Agent-World or AgentWorld comparison requests, cross-suite agent benchmark design, Tau2 comparator/profile diagnosis, paired thin-runtime-versus-GEODE experiments, mean-accuracy@8 replication, or benchmark comparability and artifact reviews.

日本語の概要は準備中です。原文の説明を表示しています。

mangowhoiscloud/geode152026年10月12日 更新

Verification checklist to prevent fake success. Detects test deletion/disabling, coverage regression, lint bypass, secret exposure. Triggered by "deception", "fake" ("가짜"), "fake success", "verification" ("검증"), "checklist" ("체크리스트"), "deletion detection" ("삭제 탐지"), "regression" keywords.

日本語の概要は準備中です。原文の説明を表示しています。

mangowhoiscloud/geode152026年10月12日 更新

Implement proven backend architecture patterns including Clean Architecture, Hexagonal Architecture, and Domain-Driven Design. Use when architecting complex backend systems or refactoring existing applications for better maintainability.

日本語の概要は準備中です。原文の説明を表示しています。

mangowhoiscloud/geode152026年10月12日 更新

Auto-search and summarize latest AI/agent papers.

日本語の概要は準備中です。原文の説明を表示しています。

mangowhoiscloud/geode152026年10月12日 更新

mangowhoiscloud のスキルをすべて見る

このスキルの問題を報告する