Skip to main content

EXTENSIONS

Selection
58 results
label:homelab

Frigate Timelapse

@mgreten/frigate-timelapse · v2026.11.10.1

Construction site timelapse capture and compilation via the Frigate NVR API. Captures periodic snapshots from any Frigate camera on a remote host via SSH, organises them into named phases, and compiles per-day and full-phase H.264 videos using ffmpeg inside the Frigate container.

upd Nov 1037 pullsA100/100

Traefik

@sntxrr/traefik · v2026.11.10.1

Keep Traefik reverse proxies, the images beside them (CrowdSec and others) and every plugin they load current, through pull requests to the repository each proxy is deployed from -- never by touching a running proxy. `inspect` reads the live proxy over SSH or its API: running version, router and middleware status, and whether plugins failed to load since this start. `check` compares the repository's pins with what is running and with upstream releases, and reports what is behind plus how the deployment is exposed: plugins downloaded from the catalog at every start (a failed download disables every plugin, and with them every router that uses one), pins that never deployed, an insecure API published on a host port, routers that are down. `bump`, dry-run by default, opens one pull request per component: an image tag, a remote plugin's version, or a re-vendored local plugin with its commit and tarball checksum recorded. It defers to an open Dependabot PR that already reaches the target, names one that falls short as superseded, refuses major upgrades unless asked, can arm GitHub auto-merge on patch or minor bumps once the base branch requires a named check (never on a major), and refuses deployments that still download plugins at startup unless asked. Plugin middleware configuration, which can carry a CrowdSec LAPI key, is never read into a resource, and neither is a plugin middleware's error text or a partial API response; the container's environment is never read. Unknown is never current: health is null while routers or the plugin-failure log could not be read, and an uncomparable version is a finding.

upd Nov 1028 pullsA100/100

Tv Curator

@lint/tv-curator · v2026.11.09.1

TV show keep-score engine — fuses Sonarr + Seerr + Tautulli + Plex into a per-series keep score with an endedUnwatched penalty for never-played completed series.

upd Nov 948 pullsB87/100

Media Curator

@lint/media-curator · v2026.11.09.1

Movie keep-score engine — fuses Radarr + Seerr + Tautulli + Plex inputs into a per-movie keep score with reason breakdown; emits drop_candidates for downstream pruning.

upd Nov 951 pullsB87/100

Hermes Config

@mgreten/hermes-config · v2026.11.06.1

Manage Hermes AI agent model configuration — set the main model, configure auxiliary task overrides, and restart the gateway. Safe edit protocol: backup → edit → validate → restart.

upd Nov 655 pullsA100/100

Swamp Go Brr

@magistr/swamp-go-brr · v2026.10.25.10

An autonomous, driver-free development loop for coding agents. You hand it an

upd Oct 2544 pullsA100/100

Pihole

@magistr/pihole · v2026.10.25.10

Pi-hole custom DNS record management for swamp — list, add, delete, and

upd Oct 25144 pullsA100/100

Libvirt

@bad-at-naming/libvirt · v2026.10.25.10

Comprehensive libvirt/virsh management — VMs, networks, storage pools,

upd Oct 2563 pullsA100/100

Firecracker

@magistr/firecracker · v2026.10.25.10

Firecracker microVM lifecycle management via SSH + the Unix-socket REST API.

upd Oct 2584 pullsA100/100

Diskio

@magistr/diskio · v2026.10.25.10

Answer "what is actually reading this disk?" on a Linux host whose storage

upd Oct 2541 pullsA100/100

Uptime Kuma

@sntxrr/uptime-kuma · v2026.10.25.1

Read-only Uptime Kuma monitor history for triage: the last N heartbeats of one monitor -- status, time, message, ping -- plus a summary of the current status, since when, and how many times it changed in the window. Kuma has no REST endpoint for heartbeat history, and its socket.io API needs the instance's only kind of login, an administrator, so any credential that can read history can also delete monitors and read notifier secrets. This model instead reads kuma.db over SSH with the host's python3, opening the database mode=ro with PRAGMA query_only=ON so SQLite itself refuses any write, and selects named columns from the monitor and heartbeat tables only: never notifier settings, credentials, monitor URLs or hostnames, or heartbeat response bodies. Monitors are found by id or by name (exact, else a unique case-insensitive substring; ambiguity is an error listing the candidates). It needs no Kuma login, never enters the container, never copies the database off the host, and has no write method.

upd Oct 2519 pullsA100/100

Docker Ps

@sntxrr/docker-ps · v2026.10.25.1

A read-only container listing of a Docker host over SSH, narrow enough to hand to an unattended triage agent. It answers the first question an alert usually raises -- what is running on that host, and which container holds that port -- with name, image, state, status, published ports and creation time per container. It runs `docker ps` with an explicit, per-field `--format` template and rebuilds each row from those fields alone, so no environment variable and no label value ever leaves the host: `docker inspect` returns every container's environment, which is where compose files keep database passwords and API tokens, and `docker ps --format '{{json .}}'` includes the Labels field. The type has exactly one method and no free-form command or format argument, so a model instance exposes no write verb at all. Host key checking stays on by default.

upd Oct 2522 pullsA100/100

Log Tail

@sntxrr/log-tail · v2026.10.25.1

Read the last lines of a log chosen by NAME from an allowlist fixed in the model definition, with secret values redacted. Built for an alert-triage agent that has no shell: an alert names a log, the cause is usually one line near its end, and the alternatives are a human tailing it for the agent or a general file-read tool that can read anything. The caller passes a name, never a path; the name-to-path map lives in the definition, where only its owner can change it. A name is matched with Object.hasOwn and a name that is not name-shaped is never echoed, because `--input name=@file` would otherwise turn a refusal into a read. The configured path must be absolute, normalised, a regular file, not a symlink and equal to its own realpath, and the opened file's device and inode must match the lstat taken before the open, so a symlink swapped in between is refused rather than followed. Output is capped at 200 lines. Redaction drops values and keeps names (`password=[REDACTED]`) across key=value, JSON, YAML, CLI flags, URL userinfo, Authorization headers and bare token shapes, and fails closed: a line mentioning a password manager CLI or a private key is withheld whole, a line still credential-shaped after redaction is withheld whole, and an error in redaction fails the method. Both methods are read-only.

upd Oct 2518 pullsA100/100

Homelab

@me__zlxp/homelab · v2026.10.24.1

Homelab health as swamp data: TrueNAS SCALE pools, apps and alerts over its JSON-RPC API with certificate pinning, Immich library and job queues, and go2rtc camera streams checked frame by frame.

upd Oct 247 pullsB87/100

Docker Reclaim

@sntxrr/docker-reclaim · v2026.10.24.2

Measure what a Docker host's disk is actually spent on, plan what can be freed without destroying a rollback target, and free it. The command people reach for is the wrong one: `docker image prune` — dangling or `-a` — never touches the BUILD CACHE, and on a host that builds its own images the build cache is usually most of the footprint. On the host this was written against it was 59.5GB of a 77GB reclaim, which no amount of image pruning would have recovered. Three behaviours are deliberate because each one corresponds to a way the obvious command does damage or lies. `docker image prune -a` removes every image no container references, which is precisely the set of previous versions a rollback needs; this model instead protects the newest `keepVersions` tags of each repository and removes specific image IDs, so reclaiming space does not quietly cost you the ability to go back. `docker builder prune -af` deletes the cache that makes the next build fast, and cache that is minutes old is the most valuable cache there is; pruning is therefore age-filtered via `--filter until=`, so a scheduled run takes cold layers and leaves the working set. The prune runs with `-a` and the age filter as its guard, because without `-a` BuildKit leaves out records shared with the image store, and on a Docker engine an age-filtered prune then frees almost nothing. The prune names the engine's own builder (`--builder "$(docker context show)"`) rather than whichever buildx builder is current, and an `engine-builder` pre-flight check refuses `reclaim` unless that builder uses the docker driver, so cache is never pruned on an engine the model did not measure. And Docker reports sizes as rounded human strings where 1000 versus 1024 is a real 7% at terabyte scale, so parsing is explicit and unit-aware and the freed figure is measured by observing free space before and after rather than by summing estimates. Each method writes its own data instance — `usage`, `plan`, `reclaim` — so an expression always gets the schema it expects rather than whichever method ran last. `observe` and `plan` are read-only and safe to schedule at any frequency; `reclaim` is dry-run unless passed `apply: true`, and even then never removes an image that any container references, running or stopped. Host key checking is left ON by default — `StrictHostKeyChecking=no` is available but is not the default, because a model whose entire purpose is running destructive commands as a privileged user is the last place to accept an unverified host key. For a Docker engine on the machine swamp itself runs on, such as Docker Desktop on a workstation, set `sshHost: local`: each command runs through `sh -c` as the invoking user, with no SSH server required, and the commands and protections are identical to the SSH path.

upd Oct 2497 pullsA100/100

Vikunja Kanban

@sntxrr/vikunja-kanban · v2026.10.23.1

Vikunja kanban orchestrator — creates tasks in a Vikunja project directly

upd Oct 23127 pullsA100/100

Jetkvm

@sntxrr/jetkvm · v2026.10.19.2

Operate a JetKVM from swamp: health checks and Tailscale updates. The JetKVM web UI does almost everything over a WebRTC data channel, so video, USB, version and settings are not on HTTP and there is no snapshot endpoint; this model reads the unauthenticated /metrics page for the app version and NTP state, and everything else over SSH as root (developer mode, key auth, host key pinned or already known; never trust-on-first-use). health flags app and system firmware updates (via JetKVM's release API), USB gadget not enumerated, low disk, clock skew, and Tailscale down, offline, out of date or missing its boot script; with the optional tailnetIp it also fetches /metrics over the tailnet, which catches a node deleted from the tailnet that tailscaled still reports as Running. Reads and writes are separate methods: tailscale reports status or plans an update and never changes the device; tailscaleUpdate always acts, behind labeled live pre-flight checks (installed, boot script, free space, tailnet path, no swap already running). The device has no CA store, so `tailscale update` cannot run on it: tailscaleUpdate downloads the official arm tarball on the swamp host, checks its sha256, streams it over SSH, backs up the running binaries, swaps them in the layout `tailscale configure jetkvm` expects, restarts tailscaled without a reboot, and rolls back by itself unless the node rejoins Running and online with the same tailnet IP (when the old one was readable) and, with tailnetIp, answers over the tailnet from the swamp host. Every update outcome is recorded as its own update instance, so a later status run never hides a failed update.

upd Oct 1945 pullsA100/100

Glinet Kvm

@sntxrr/glinet-kvm · v2026.10.19.1

Operate GL.iNet KVMs (Comet GL-RM1 / GL-RM1PE) from swamp: health, screenshots, ATX power and virtual media. The firmware is a fork of PiKVM's kvmd, and three of its behaviours shape this model. The factory TLS certificate is self-signed, names CN=localhost and expired in 1979, so it cannot be validated or pinned; HTTP goes through curl and the viewer through openssl, with credentials on stdin rather than argv. The video streamer sleeps until a viewer connects, so a bare snapshot request answers 503 whether or not there is a signal; screenshot and health open a receive-only viewer session first and never send input. Without the ATX add-on board, the API still reports power 'off' for a running machine; this model reports the state as unknown and refuses to press a button wired to nothing. health flags available firmware updates, missing HDMI signal, offline USB HID, forgotten virtual media that could hijack the next boot, clock skew, Tailscale (the device's own state, plus an optional probe of its tailnet IP that catches a node deleted from the tailnet, which the device still reports as Running), and security posture (2FA off, browser root shell on, untrusted TLS). Reads and writes are separate methods: atx and msd report state or record a plan and never change the device; atxPress and msdChange act, behind labeled live pre-flight checks (ATX board present and not busy; virtual media enabled and not busy).

upd Oct 1980 pullsA100/100

Technitium

@jamesakeech/technitium · v2026.10.15.1

Manage a Technitium DNS server or cluster: zones, records, block/allow lists, cache, diagnostics, settings backup and restore, and cluster membership. Written after rebuilding a live authoritative pair, and shaped by what that exposed. `record_ensure` is an idempotent upsert Technitium's own API cannot express — add fails on an existing record, update fails on an absent one — and it works for every record type including TLSA, because the read and write field names for TLSA, SSHFP and URI are not the same names and are mapped per type rather than renamed globally. `settings_backup` includes every section by default, so it produces a backup a server can actually be rebuilt from. `cluster_init`, `cluster_join` and `cluster_state` cover the membership a settings backup can never carry, alongside the TLS listener and certificate it also omits — the three things that make a restored node look broken.

upd Oct 1533 pullsA93/100

Pocket Id

@jamesakeech/pocket-id · v2026.10.15.1

Observability for a Pocket ID instance — the passkey-based OIDC provider. `health` is a cheap, admin-free probe that separates an unreachable host from a wrong API key from a key whose owner is not an admin, and reports version drift; it writes its result as data rather than throwing, so a failed run still leaves a truthful record. `sync` fans out one resource per user, OIDC client, group and API key, joins each user against their passkeys and their sign-ins and each client against its authorizations, then scores the whole instance into `instance.findings`: accounts with no passkey that therefore cannot sign in, public clients with PKCE disabled, clients nobody has authorized, empty groups, and the API key whose expiry will silently stop the sync. `syncActivity` reads a bounded window of the audit log on its own — Pocket ID has no date filter, so the window is applied by walking newest-first and stopping early, which makes a short window genuinely cheap. Read-only throughout.

upd Oct 1529 pullsA93/100

Nginx Proxy Manager

@jamesakeech/nginx-proxy-manager · v2026.10.15.1

Full lifecycle management of an Nginx Proxy Manager instance. `sync` fans out one resource per proxy host, redirection host, dead host, stream, access list and certificate, plus an instance rollup that flags expiring certificates, plain-HTTP hosts and domains claimed by more than one host. The `apply*` methods are idempotent — they match an existing object by its natural key (domain set, listening port, access list name) and update it in place, so re-running a workflow converges instead of accumulating duplicates. `setEnabled` and `delete` dispatch on an object kind and take a list of ids, so a batch is one fan-out call rather than a run per id. Let's Encrypt certificates can be requested over HTTP-01 or DNS-01 (wildcards included), renewed, or replaced with uploaded PEM files.

upd Oct 1580 pullsA93/100

Hermes Env

@mgreten/hermes-env · v2026.10.14.2

Manage Hermes agent environment variables via SSH or locally from inside the container. Supports setEnv, removeEnv, listKeys, and restartGateway with a localMode flag for self-healing from within Hermes itself.

upd Oct 1445 pullsB86/100

Unifi Fabric

@sntxrr/unifi-fabric · v2026.10.14.1

Structural health monitoring for a UniFi fabric. The `@sntxrr/unifi-fabric/topology` model's `check` method compares a declared topology against live `/stat/device` rows and reports the failures that outcome-based monitoring cannot see: a device expected on the wire that has silently fallen back to a wireless mesh uplink, attachment to the wrong upstream device, links negotiated below their expected speed, ports carrying error counters, and — the one with no equivalent elsewhere — ports that are down but have carried real traffic before, which identifies a run that used to work. The first three are detectors and gate `inSync`; error counters and dark ports are cumulative totals rather than rates, so both are reported as diagnostics and deliberately do not, since a single snapshot cannot tell a severed run from a client that went to sleep. An access point that loses its wired uplink does not fail; it meshes, keeps serving clients, and every uptime check stays green while latency quietly goes from sub-millisecond to tens of milliseconds and jittery. `uplink.type` flipping from `wire` to `wireless` is a boolean, so it is asserted exactly rather than thresholded. Read-only: never writes to the controller. Emits a flat Prometheus-ready metric series alongside the verdict, including for healthy devices, so alerts can fire on a series dropping to zero rather than on a document changing shape. Authenticates with an API key over `X-API-KEY`, which sidesteps the HTTP 499 that MFA-enabled SSO accounts return for password logins.

upd Oct 1460 pullsA100/100

Unifi Dhcp Reservation

@sntxrr/unifi-dhcp-reservation · v2026.10.14.1

Declarative DHCP fixed-IP reservations on a local UniFi controller (UDM / UDM Pro / UDM SE). `sync` reads every reservation the controller holds; `drift` compares a desired set against it without writing, reporting missing, mismatched, unmanaged, duplicate and DHCP-pool-overlapping entries; `apply` reconciles the controller to the desired set and supports `dryRun`. `device_drift` does the same read-only comparison for adopted hardware, which cannot hold reservations at all and is addressed through device config instead — the one surface `drift` is blind to. Reservations live on the legacy Network API as `user` objects carrying `use_fixedip`/`fixed_ip`. Adds TOTP/MFA login support, which the upstream @mgreten/unifi auth flow lacks — UniFi SSO accounts with MFA reject password-only logins with MFA_AUTH_REQUIRED. Catches the failure mode where a reservation silently never takes effect because its address is already claimed by a statically-configured host. `forget_client` prunes stale client records the controller still remembers, refusing any MAC that holds a reservation or a live lease unless forced.

upd Oct 1436 pullsA100/100