with one click
Skill: task-queue
npx skills add https://github.com/JansenAnalytics/claudex --skill task-queueCopy and paste this command into Claude Code to install the skill
Skill: task-queue
npx skills add https://github.com/JansenAnalytics/claudex --skill task-queueCopy and paste this command into Claude Code to install the skill
| name | task-queue |
| description | Skill: task-queue |
A persistent, JSON-backed job queue. Lets Kite schedule and manage shell commands without raw cron. One-shot tasks (run at a specific time) and recurring tasks (cron schedule) are both supported.
| File | Purpose |
|---|---|
${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/queue.json | Task store (source of truth) |
${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/cli.cjs | CLI tool for managing tasks |
${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/runner.cjs | Runner — executes due tasks |
${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/runner.log | Runner output log |
Skill copies live in ${CLAUDE_SKILLS_DIR:-$HOME/.claude-agent/.claude/skills}/task-queue/scripts/.
# Add one-shot task
node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/cli.cjs add "name" "command" --at "2026-02-21T18:00:00"
# Add recurring task (every 15 min)
node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/cli.cjs add "name" "command" --every "*/15 * * * *"
# Add immediate task (runs next runner tick)
node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/cli.cjs add "name" "command"
# List all / filter
node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/cli.cjs list
node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/cli.cjs list --status pending
# Details for one task
node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/cli.cjs status <id>
# Cancel / mark done
node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/cli.cjs cancel <id>
node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/cli.cjs done <id>
# Housekeeping
node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/cli.cjs clear --done
node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/cli.cjs clear --failed
node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/cli.cjs clear --all
# Run runner manually (normally cron does this)
node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/runner.cjs
# Tail log
tail -f ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/runner.log
* * * * * /usr/bin/node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/runner.cjs >> ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/runner.log 2>&1
{
"id": "a1b2c3d4",
"name": "human name",
"command": "echo hello",
"type": "one-shot | recurring",
"status": "pending | running | done | failed | cancelled",
"scheduled_at": "2026-02-21T18:00:00.000Z",
"cron_expr": "*/15 * * * *",
"created_at": "...",
"started_at": "...",
"finished_at": "...",
"last_run_at": "...",
"output": "...",
"error": "..."
}
*, */N, N-M, N,M,… for each fieldpending after each run so they fire again next cycledone (or failed) after their single executionPipelines are multi-step tasks where each step is checkpointed. On failure, retry skips already-completed steps and resumes from where it left off.
| File | Purpose |
|---|---|
${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/pipelines.json | Pipeline store |
${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/pipeline-runner.cjs | Cron runner for pipelines (every minute) |
${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/pipeline-runner.log | Pipeline runner log |
# Add a pipeline (steps as inline JSON)
node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/cli.cjs pipeline add "Deploy app" \
--steps '[{"name":"Run tests","command":"cd ~/myapp && npm test"},{"name":"Push","command":"cd ~/myapp && git push"}]'
# Add a pipeline from a steps file
node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/cli.cjs pipeline add "Deploy app" --steps-file /path/to/steps.json
# List all pipelines (shows step progress)
node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/cli.cjs pipeline list
# Detailed status (per-step output and errors)
node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/cli.cjs pipeline status <id>
# Retry a failed pipeline (skips done steps, re-runs from failure point)
node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/cli.cjs pipeline retry <id>
# Cancel / delete
node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/cli.cjs pipeline cancel <id>
node ${TASK_QUEUE_HOME:-$HOME/projects/task-queue}/cli.cjs pipeline delete <id>
[
{ "name": "Step 1", "command": "echo hello", "timeout_ms": 60000 },
{ "name": "Step 2", "command": "echo world" }
]
{
"id": "abc12345",
"name": "My pipeline",
"status": "queued | running | done | failed | cancelled",
"current_step": 1,
"steps": [
{
"id": "step-1",
"name": "Step name",
"command": "shell command",
"status": "pending | running | done | failed | skipped",
"timeout_ms": 600000,
"started_at": null,
"finished_at": null,
"output": null,
"error": null
}
]
}
Semantic memory search across all agent memories and conversation history. Use BEFORE answering questions about prior work, decisions, dates, people, preferences, projects, or past conversations. Also use when asked "do you remember", "what did we discuss", "when did we", etc.
Get current weather and forecasts. Use when the user asks about weather, temperature, or forecasts for any location.
Set up and use 1Password CLI (op). Use when installing the CLI, enabling desktop app integration, signing in (single or multi-account), or reading/injecting/running secrets via op.
Accessibility auditing: WCAG compliance checking, contrast ratios, ARIA labels, keyboard navigation, semantic HTML, screen reader compatibility.
ADR Manager Skill
Autonomous API testing and evaluation. Tests any REST API for correctness, security, performance, error handling, and standards compliance. Discovers endpoints, probes with valid/invalid/edge-case payloads, checks auth, response times, injection vulnerabilities, and generates severity-scored reports with actionable fixes. Use before any API "done" claim.