agentleFS
WhispersOfJWhispersOfJ/arrfleet-toolkit/skills/download-client-orchestrator/SKILL.md

download-client-orchestrator

Skill1 starsChanged 4 months agoMITView on GitHub

Inspect and manage a download client's queue and health — any SABnzbd-API-compatible client (SABnzbd, NZBGet, NzbDAV) or qBittorrent-Web-API-compatible torrent client wired into an Arr stack. Use when the user asks about stuck or failed downloads, wants to see what's queued, needs to clear a jammed queue item, or wants to confirm the download client is reachable by the Arr apps. Trigger phrases: "what's stuck in the download queue", "check sabnzbd status", "clear failed downloads", "why isn't radarr grabbing anything", "check qbittorrent queue".

  • Reads credentials

The file

Files
SKILL.md

Download Client Orchestrator

Queries a download client's own API directly (bypassing the Arr apps) to answer "is the download client itself the problem?" before assuming it's an indexer, mount, or Arr config issue. Complements health-monitor (container-level health) by checking queue-level health: stuck items, failed items, and reachability.

Supported clients

Built-in defaults cover the common SABnzbd-API-compatible Usenet clients (SABnzbd, NZBGet in its SABnzbd-emulation mode, NzbDAV) and qBittorrent's Web API for torrent clients. Add a project-specific client — a different port, a proxy path prefix, a fork — via .arrfleet-clients.json next to docker-compose.yml:

{"my-client": {"port": 8081, "path": "/sabnzbd", "kind": "sabnzbd"}}

kind is "sabnzbd" or "qbittorrent"; path is the base path the client's API is actually mounted under (a stock install usually expects /api directly, but some clients mount it behind a prefix — check that client's own docs before guessing).

Auth / config

<CLIENT>_URL / <CLIENT>_API_KEY     e.g. SABNZBD_URL / SABNZBD_API_KEY

<CLIENT> is the client name uppercased (hyphens become underscores). Defaults to http://localhost:<port> if unset. Never hardcode a real LAN IP or credential as a fallback — env vars only, localhost default.

Usage

python3 orchestrator.py queue sabnzbd                 # list current queue items + state
python3 orchestrator.py failed sabnzbd                 # list only failed/error items
python3 orchestrator.py retry sabnzbd <item-id>        # re-queue a failed item
python3 orchestrator.py clear-failed sabnzbd           # remove all failed items (asks first)
python3 orchestrator.py reachability                   # ping every configured client, report up/down

Interpreting results

  • A client that's unreachable (reachability fails) almost always means the container is down or mid-restart — hand off to docker-compose-manager status <client> rather than digging further here.
  • A client that's reachable but has a growing "failed" queue with the same error message repeated is usually an indexer/API-limit problem upstream (e.g. in Prowlarr), not this skill's job to fix — report the pattern, don't guess at indexer config.
  • Stuck-but-not-failed items (state unchanged across repeated queue calls) often correlate with a stale download-path mount — cross-check with docker-compose-manager cascade-map before assuming the download client itself is broken.

Safety rules

  • clear-failed prompts for confirmation — it's a bulk destructive action on queue state.
  • retry only touches the single item ID given; never retries an entire queue implicitly.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.