plex-mcp
Skill2 starsChanged 9 months agoMITView on GitHub
Manage a Plex Media Server via 22 portmanteau MCP tools (libraries, media, search, streaming, users, playlists, metadata, performance, reports, RAG, taste, enrichment, arr, lifecycle). Use when the task touches Plex libraries, playback, users, or media metadata.
In this repository8 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- .cursor/rules/architect.mdcCursor rule
- llms-full.txtllms.txt
- llms.txtllms.txt
- full-stack-developerSkill
- mcp-server-architectSkill
- plex-mcp-roadmapSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 57
Search → plex_media get_details (confirm the wrong match) → plex_metadata match (candidate hits) → fix_match → refresh → get_details again …
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
PlexMCP skill
Talk to a Plex Media Server through portmanteau tools — one tool per domain,
each with an operation parameter. Never invent identifiers, URLs, or tokens;
discover them first.
Session startup
plex_server(operation="status")— reachable? version? load?plex_library(operation="list")— sections, types, sizes.plex_rag(operation="status")— only when the task needs semantic search.
Tool surface (24 tools)
| Tool | Key operations |
plex_library | list, get, list_items, create, scan, refresh, optimize, empty_trash (!), add/remove_location, delete (!) |
plex_media | browse, search, get_details, get_recent, get_stream_url, update_metadata |
plex_search | search, advanced_search, suggest, recent_searches, save_search |
plex_streaming | list_sessions, list_clients, play, pause, stop, seek, control |
plex_server | status, info, health, maintenance, restart (!), update (!) |
plex_user | list, get, create, update, delete (!), update_permissions |
plex_playlist | list, get, create, update, delete (!), add/remove_items, get_analytics |
plex_collections | list, get, create, update, delete (!), add/remove_items |
plex_metadata | analyze, match, fix_match, refresh, refresh_all, update, organize |
plex_organization | analyze, organize, fix_issues, clean_bundles, optimize_database |
plex_performance | transcode getters before setters; set_quality, throttling, profiles |
plex_quality | list/get/create/update/delete profiles, set_default |
plex_reporting | library_stats, content_report, user_activity, usage/performance reports, export_report (json/csv/html) |
plex_rag | status, sync_metadata, sync_subtitles, semantic_search, search_subtitles |
plex_taste | status, rebuild_from_history, sync, profile, score_item, score_library |
plex_media_enrichment | enrich, bulk_enrich, get_enrichment, get_related, search_by_theme/person, stale_report |
plex_integration | list_integrations, vienna_recommendations, european_content, anime_season_info, configure, sync |
plex_audio_mgr | list/select streams, volume, mute, handover |
plex_ffmpeg_mgr | probe first, then sync/extract operations, revert |
arr_stack | status (read-only Radarr/Sonarr/Lidarr snapshot) |
plex_help | help, list_tools, tool_info, examples, api_docs |
plex_shutdown | status, shutdown (!) |
agentic_plex_workflow | orchestrate (needs sampling or PLEX_SAMPLING_BASE_URL) |
plex_natural_assistant | answer (no tools) |
(!) needs explicit user confirmation (destructive or interruptive).
Workflow 1 — triage an unknown server
plex_server status → plex_library list → summarize versions, sections,
anything offline. Unreachable: check URL/token via the webapp Settings PIN
flow, do not retry blindly.
Workflow 2 — fix a bad metadata match
Search → plex_media get_details (confirm the wrong match) →
plex_metadata match (candidate hits) → fix_match → refresh →
get_details again to verify art and cast.
Workflow 3 — semantic discovery
plex_rag status → if empty, sync_metadata (warn: minutes to hours) →
semantic_search in the user's words → search_subtitles for quoted
dialogue. Keyword search never needs the index.
Safety
Confirm before: empty_trash, delete (library/playlist/user), remove_location, PMS restart/update, killing others' sessions, server shutdown. Reads and searches never need confirmation. Tokens live in env/keystore — use the webapp PIN flow, never paste secrets. Verify every write with its paired read (update→get_details, add_items→get, setter→getter→live session).
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
Reviews can't be read right now.
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.