Add the starpunk-mcp container (same image, uvicorn :8001, shared env via
anchor, BACKGROUND_WORKERS_ENABLED=false, socket healthcheck) and route
/mcp and /.well-known/oauth-protected-resource to it in Caddy. Bump
starpunk_version to v2.15.0 (needs the v2.15.0 tag built first).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Mastodon allows ~99MB video uploads; the 40MiB default dropped many
remote videos with 'file size limit reached'.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Podman reports the sandbox image as docker.io/nousresearch/...; Hermes
compares image names as text, so the short name made it delete and
recreate the sandbox on every reuse. Use the full name.
Hermes 0.23.1 looks for Chromium in chrome-linux/, but the ARM64 image
ships chrome-linux-arm64/ (hermes-agent #153626), so browser tools said
there was no Chromium. Mount a folder with a chrome-linux/chrome script
that runs the image's Chromium. Tested in a fresh sandbox: the probe
finds it and browser_exec loads pages.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
terminal.cwd "." resolved against each process's working directory, so
the dashboard's sandbox mounted / (the whole disk, writable) and the
gateway's mounted ~/.hermes (including .env). Point it at ~/hermes.
The cua-driver symlink in ~/.hermes/skills made Hermes give each sandbox
a temp copy of the skills and delete the previous copy, emptying running
sandboxes' skills (hermes-agent #53630). Read Cua's pack through
skills.external_dirs and remove the symlink.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Hermes's OpenRouter code treats Claude models it doesn't list as
reasoning-mandatory and drops the guardian's "reasoning off" for Haiku
5.5, which then sometimes spends the 16-token budget thinking and
returns nothing, escalating the command to the user. Send
extra_body.reasoning {effort: none} for the approval task; replays of
the failing commands went 21/21 correct.
Also create ~/hermes, the folder the Hermex Bot Chat session works in.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Hermes's built-in web search was the free Brave API, which returns 429s
under agent bursts; real-user reports (Hermes #80201, PR #36838, HN) also
rule out self-hosted SearXNG for always-on use. The user has Kagi, so:
- Kagi's own hosted MCP server (https://mcp.kagi.com/mcp, Bearer API key,
paid API with a monthly cap set in Kagi). Its two tools,
kagi_search_fetch and kagi_extract, are marked read-only, so
trust: untrusted never asks.
- agent.disabled_toolsets gains `web`, removing the built-in
web_search/web_extract everywhere, so Kagi is the only search path.
- KAGI_API_KEY from the vault into ~/.hermes/.env.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The smart-approval guardian asks for a one-word verdict with reasoning off
and max_tokens=16 (title generation: 64). GLM-5.3 Flash's reasoning is
mandatory on OpenRouter, so Hermes floors it to effort=low, the model
spends the budget thinking and returns nothing (finish_reason=length), and
every flagged command escalated to the user. Move approval, titles, memory
query rewrite and MCP sampling to Claude Haiku 5.5 (reasoning off by
default); compression, curator and the mail monitor stay on GLM.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
`git worktree list --porcelain | awk '...{print $2; exit}'` exits after
the first line; with 27 worktrees git is still writing, gets SIGPIPE, and
under `set -o pipefail` every bin/ helper (deploy, vault-run, worktree)
exited 141 with no output. Let awk read to the end instead.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
`hermes config get` fills in and masks ${VAR} references, so the Karakeep
MCP entry (env KARAKEEP_API_KEY: ${KARAKEEP_API_KEY}) never matched and
every deploy re-set it and restarted the gateway. It also prints defaults
for unset keys and exits 1 for keys without one. Compare against
~/.hermes/config.yaml itself instead.
`hermes gateway restart` without a terminal refuses while anything is
running unless given --yes (docs, reference/cli-commands); the failed
handler left the gateway restarting under launchd but the play failed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Hermes can search and read bookmarks, create them, tag them and add them
to lists from Telegram. Per Karakeep's MCP README and Hermes's MCP docs:
- the official server image (pinned 0.34.1) runs in a throwaway Podman
container (`podman run -i --rm`), started on first use and stopped
after 10 idle minutes;
- only search/read/create/tag/add-to-list tools are registered (no update
or delete); trust: full because Karakeep doesn't mark its search tools
read-only;
- the API key is in the vault, written to ~/.hermes/.env via the new
hermes_env_secrets, and passed into the container by name.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Models, from the Nov 2026 review (Nous Hermes Index, OpenRouter data):
- chat: openai/gpt-6.1-sol (user's pick), reasoning high (was ultra)
- side tasks (compression, approval, titles, mcp, memory query rewrite,
curator, mail monitor) and unpinned cron: z-ai/glm-5.3-flash
- subagents: deepseek/deepseek-v4.1-flash
- /model aliases: deepseek, haiku
agent.reasoning_effort needs --force (hermes-agent #140221), so settings
can list keys in hermes_settings_force. `config get` exits 1 for unset
keys without a default; that now reads as unset, not a failure.
hermes-model-review: monthly cron job on Sol (pinned, high effort,
continuity) that redoes this research and sends keep/change
recommendations as Ansible keys to Telegram. A script injects the live
model setup and 30-day usage into its prompt. hermes_cron_jobs gains
model/provider, reasoning_effort and continuity.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
bad2426 failed silently: over SSH `sysadminctl -autologin set` can't reach
the loginwindow session agent and fails with SACSetAutoLoginPassword
error:22 (rc 0), leaving autoLoginUser set without /etc/kcpassword. Per
gulasz101/homelab-2nd#178 (measured 2026-11-09), run it inside the console
user's session with `launchctl asuser <uid>`, password on stdin. It accepts
wrong passwords, so check the vault password with `dscl -authonly` first
(as brunerd's setAutoLogin.sh does). Idempotency now checks autoLoginUser
and /etc/kcpassword instead of the status text, which never matched.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Mac powers back on after a power cut but stopped at the login window,
so nothing that needs a login session (dashboard, Podman VM for Hermes's
sandbox, BlueBubbles, Beeper) came back. Turn on automatic login with
sysadminctl, using phil's password from the vault. FileVault is off, so it
works unattended. User accepted the physical-access trade-off.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
`podman machine start` exits once the VM is up, and launchd then killed the
VM processes left in the job's process group, so a VM started by the job
stopped ~20 s later. Keep them with AbandonProcessGroup, and reload the job
when its plist changes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Non-interactively, `hermes config get` prints only the default for a key
that isn't in config.yaml, so an unset key can't be detected and the play
kept trying (and failing) to unset terminal.cwd. Set it to Hermes's own
default "." instead and drop the unset branch.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
`hermes config get` writes "Config key not set" to stderr and Hermes's
default to stdout, so terminal.cwd: null looked set ("."), the play tried
to unset it again and `unset` failed on the missing key.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The VM resize in a92d4c1 worked but left the VM stopped: a VM started from
an Ansible task is stopped when the task ends. Nothing started the VM at
login either, so after a reboot Hermes's sandbox stayed down.
- launchd job local.podman-machine runs `podman machine start` at login;
the resize now starts the VM through it and waits until it's running.
- Hermes setting changes also restart the dashboard, which keeps its own
copy of the settings (it kept logging the old /workspace cwd).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Via hermes_settings (now compared structurally, with null = unset, and a
gateway restart on change):
- timezone America/Denver (was unset)
- auth.adopt_external_logins false: never borrow Claude Code's login
- terminal.cwd unset: stops the /workspace warnings (#140527)
- agent.disabled_toolsets [kanban]: unused, 25 KB of tool schema per call
- skills.write_approval true
- approvals.deny: force-push, curl|sh, dd to devices (docs' examples)
- streaming.enabled true, stt.local.model small
- terminal.container_memory 3072, inside a 4 GB Podman VM (was 2 GB VM
with a 5 GB container limit), set by a new podman task
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Hermes has no auto-update for itself, and /update from Telegram can leave
a launchd gateway offline. Instead, a no-agent Hermes cron job runs
`hermes update --check` on Mondays at 09:07 and messages Telegram only when
a new stable release is out (a failing check sends an error alert).
Adds hermes_cron_jobs (scripts in files/hermes-scripts/, copied to
~/.hermes/scripts/; jobs created when missing) so the deal checker can
follow the same pattern.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Adds a small hermes_settings map that Ansible checks with
`hermes config get` and fixes with `hermes config set`, leaving the rest
of config.yaml to Hermes (its updates migrate that file).
First setting: updates.pre_update_backup: full (was false), so every
`hermes update` zips ~/.hermes first. Rollback isn't supported once a new
version has written state.db, so the backup is the way back.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The dashboard used to be started by hand as `serve --host 0.0.0.0`, so it
was reachable from the LAN and didn't survive a reboot or a Hermes update.
Now, following Hermes's docs:
- launchd agent local.hermes-dashboard runs `hermes dashboard` on
127.0.0.1:9119 (KeepAlive, --skip-build since `hermes update` builds it).
- Tailscale Serve publishes it on https://cogitator-02.porgy-porgy.ts.net,
tailnet only. No module exists for Serve; the task reads
`serve status --json` and only sets it when it differs.
- Login: scrypt password hash and a real signing secret from the vault
(the old secret was the literal text "$(openssl rand -base64 32)"); the
plaintext password line is removed from .env since it overrides the hash.
- After Serve is set, the play checks /api/status demands the basic login;
if not, it turns Serve off and fails.
- Hermes's Python no longer needs a firewall exception, so that's gone.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- pmset: never sleep, no disk sleep, restart after a power cut, wake on
network. The Mac had sleep 1 and only stayed up because BlueBubbles runs
caffeinate.
- Application firewall on, keeping signed software allowed automatically,
stealth and block-all off. Hermes's ad hoc signed Python (dashboard on
9119) is allowed explicitly, since an unattended Mac blocks unsigned
programs silently.
- Local id_ed25519.pub in phil's authorized_keys (already there).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
cogitator-02 is an Apple Silicon Mac on the tailnet running Hermes Agent.
It gets a `mac` group and mac.yml, kept apart from site.yml so it never
runs with the server's deploys (PLAYBOOK=mac.yml bin/deploy).
The mac role starts by recording the Homebrew taps, formulae and casks
already installed (present only, never upgrades or removes), so the first
run changes nothing.
No passwordless sudo on the Mac: Hermes runs as phil and takes commands
over Telegram. Ansible's sudo password is vault_cogitator02.become_password
and only tasks with become: true use it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
ckulka/baikal is stuck on Baikal 0.10.1 (PHP 8.1). dschoepel/baikal is a
drop-in successor of its nginx variant (same volumes, UID 101) with
Baikal 0.13.0 on PHP 8.4. Upgrade tested on a copy of the data.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Compose deploys used the default pull policy (missing), so switching a
service to a tag that was already cached deployed a year-old image
(paperless-ngx and heyform :latest). module_defaults now sets pull: always
for every docker_compose_v2 task.
dockcheck pulls the image it updates, but then ran compose up for the
whole stack, which could recreate other containers in the stack from
cached images, including ones it excludes. OnlySpecific=true limits it to
the updated container.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The 1.7 -> 1.22 upgrade is done; the app gets the dockcheck label.
Redis and Postgres stay excluded in dockcheck.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
1.17.0 background jobs drained cleanly. 1.18.1 (skipping 1.18.0, which
needed the hotfix) converts Areas to Places and re-detects visits across
the whole history; the changes can't be rolled back, so a fresh dump was
taken first (~/backups/dawarich-1.17.0-20261109). Redis goes to 8 as in
upstream's compose since 1.18.0.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Follows the 1.17.0 upgrade notes and upstream compose: the separate
dawarich_sidekiq service is gone (removed on the host before deploying),
the app runs Sidekiq embedded (SIDEKIQ_EMBEDDED=true, WEB_CONCURRENCY=0,
BACKGROUND_PROCESSING_CONCURRENCY=2), gets stop_grace_period 60s, and the
limits match upstream (1 CPU, 4G).
Fresh dump before this step: ~/backups/dawarich-1.16.0-20261109.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
1.14.5 background jobs fully drained (checked queues, scheduled and retry
sets, and job logs). 1.16.0 is the last release with the separate sidekiq
container and merges duplicate legacy points.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
4.4 is end-of-life. Moved by mongodump/mongorestore into a new volume
(heyform_database7); the 4.4 volume stays for rollback.
The restore also fixes a data mix-up. From Nov 2024 the compose file set
MONGO_URI='mongodb://mongo:27017/heyform' in list form, so the quotes
became part of the value and HeyForm used a database literally named
"heyform'". The v3 upgrade (a1a79af) dropped the quotes, which switched
HeyForm to the old "heyform" database from the first week of setup, so
the real Repairs and Volunteer forms (91 submissions) disappeared. The new
volume holds "heyform'" restored as "heyform". Tested first with v3 on a
throwaway Mongo 7.0: both forms load.
Backups: ~/backups/heyform-mongo4.4-20261109 on the server (dump of both
databases + tar of the 4.4 volume).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
gitea/gitea:1 stopped moving in August when Gitea renumbered to 28.0, and
1.27 gets no more fixes (28.0 fixed ~20 CVEs). Track :latest so the next
renumber can't freeze it again. The act_runner image was renamed to
gitea/runner; the old one stopped updating in April. The .runner
registration file and config format carry over unchanged.
v28 starts deleting Actions runs older than 400 days; set
RUN_RETENTION_DAYS=0 to keep the history.
Checked for this instance: ROOT_URL/SSH_DOMAIN set, no ALLOWED_HOST_LIST,
no webhooks or custom templates, one pull mirror from a public host.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Both earlier failures had one cause: the host still had old images cached
under the :latest tag (paperless-ngx from 2025-09, i.e. 2.x; heyform from
2025-02, i.e. v2), and the compose task only pulls missing images. So the
switch to :latest started year-old code. Pulled the real :latest images on
the host. Paperless 3.4.2 has no pending migrations on this database
(checked on a copy), and HeyForm v3.0.6 listens on 9157 as before.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
1.10.3 migrated cleanly and its background jobs have drained. 1.14.5
fixes the registration bypass (GHSA-x483-3xv3-vp52) and runs the
1.11-1.14 track/visit recalculation and point backfill.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- HeyForm v3.0.6 changed the default listen port from 9157 to 8000, so
Caddy got 502s. Set APP_LISTEN_PORT=9157 explicitly.
- Paperless 3.4.2 fails its migration check on this database
(InconsistentMigrationHistory, squashed 1022..1036). Rolled back to 3.4.1
until that is investigated.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
First stop is 1.10.3 (drops the legacy lat/lon columns with the 1.10.2
lock fix). Next stops: 1.14.5 (security fix), 1.16.0, 1.17.0 (sidekiq
merged into the app), 1.18.1, 1.22.0, then back to latest.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Pinned versions never got updates, because dockcheck only notices a new
image under the same tag.
- paperless-ngx: 3.4.1 -> latest (stable only; beta/dev have own tags)
- paperless redis: 7 -> 8-alpine (reads Redis 7 data as-is)
- gotenberg: 8.39.0 -> 8 (9 changed the API Paperless calls)
- tika: 3.3.1.0 -> 3.3.3.0 (no floating 3.x tag; Tika 4 breaks Paperless)
- heyform: v3.0.5 -> latest (only moved by v* release tags)
- add mag37.dockcheck.update to all of the above
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
dockcheck runs as phil (docker group) and couldn't read
/opt/stacks/gokapi (0700 root), so the labelled gokapi container could
never be auto-updated. Paperless's root-only .env (2024 leftover, unused,
same values as docker-compose.env) made docker compose fail for phil.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Install dockcheck as a git clone in ~/.local/share/dockcheck with
AutoSelfUpdate (git pull also updates notify templates and urls.list);
no pinned versions. regctl is refreshed weekly by run_dockcheck.sh.
- Since dockcheck ~v0.7, OnlyLabel skips unlabelled containers entirely, so
they got no update notices. run_dockcheck.sh now runs twice: update
labelled containers quietly, then check everything and notify once.
- Running dockcheck.sh by hand now checks everything and changes nothing.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Pin dockcheck (v0.8.3) and regctl (v0.11.7) and re-download on version
bumps; get_url never replaced existing files, and dockcheck never updates
regctl itself (server had regctl v0.8.3 from June 2025)
- New options: DaysOld=3, BackupForDays=7 (rollback), MaxAsync=10,
MonoMode for the log, PrintReleaseURL with upstream urls.list
- Drop authentik-postgresql-1 from Exclude (container is gone)
- Wrapper logged exit code 0 always ($? read after $(date)); fixed
- Rotate /var/log/dockcheck/dockcheck.log weekly
- Remove stale ~/.local/bin/notify_v2.sh
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Pass the Threads token (vault_starpunk.threads_access_token, Meta app
"starpunk") and username philskents. StarPunk refreshes the token daily
and keeps the current one in /data/threads_token.json.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The server now runs Debian 13.7. systemd 257's ssh generator logs a
harmless AF_VSOCK error on netcup VMs, so the common role masks it.
The plan file now records how the upgrade went and what's left.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Debian 13 drops apt-key and software-properties-common, so the old tasks
would fail after the upgrade. The key now lives in /etc/apt/keyrings
(same fingerprint, 9DC8...CD88), the repo is a deb822 file whose release
name follows the installed Debian, and the old .list file is removed.
The apt tasks get a docker_apt tag so they can be deployed on their own.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Browsers will use HTTPS for each site for a year. includeSubDomains is
left off so hosts not served by this Caddy are unaffected.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
v0.1.0 (Feb 2025) has dozens of published security advisories fixed in
v3. Tested against a copy of the database first; forms load unchanged.
v3 listens on 9157, keeps uploads under packages/server, and uses Redis 8
commands, so KeyDB is replaced by redis:8-alpine. The homepage URL is now
https (v3's CORS allowlist is built from it), Caddy is the trusted proxy,
public sign-up is off, and integration credentials get their own key.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
StarPunk 2.1.0 copies new notes to GoToSocial and keeps edits and
deletes in step. Token for the app "StarPunk crossposting" (scopes
write:statuses write:media) is in vault_starpunk.gotosocial_token; the
container reaches social. through Caddy (hairpin).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2.20.15 has run its migrations, so this is the supported jump to v3.
Search moves to Tantivy and is rebuilt on first start.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
v3 can only be reached from 2.20.15, so stop tracking :latest. Also set
the two settings v3 needs from us: trust Caddy as the proxy (v3's login
rate limiting), and keep rejecting duplicate documents.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
StarPunk's Gitea Actions workflow now builds and pushes the image on
vX.Y.Z tags. Pin the tag in starpunk_version (v2.0.2, the first CI build)
and always pull it, so an image built by hand on the server with the
same name can't be deployed instead.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Karakeep's Chrome was alpine-chrome 123 (April 2024) and it opens every
bookmarked URL; it now uses Karakeep's own Chrome image like upstream.
Meilisearch goes from 1.13.3 to 1.41.0 with a dumpless in-place upgrade.
Gotenberg for Paperless goes from 8.7 to 8.39.0.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
sshd accepted passwords from the internet (about 12,000 failed logins a
day) and root logins with a key. Keys only now, no root, no X11 forwarding.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The gondulf DNS record is deleted (checked against public DNS), so its
removed_records entry is no longer needed. vault_gondulf was also removed
from the vault.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Jobs run inside the runner container (label docker:host), and with the
host's Docker socket mounted every workflow had root on the server. The
socket was only there for StarPunk's image build, which has never worked
(it logged in to gitea:3000, which the host daemon can't resolve).
Image builds now go to a rootless BuildKit container that shares only a
unix socket with the runner. Site builds keep their /sites mount.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
StarPunk 2.0.0 logs in with Pocket ID and issues its own Micropub tokens
(ADR-064 in the StarPunk repo), so the Gondulf IndieAuth server is no
longer used. Removes its task, templates, Caddy route, subdomain and
service entry, and moves its DNS record to removed_records.
The StarPunk template was still on v1.6.0-rc.2 with Gondulf hairpin
hosts; deploying it would have broken login. It now matches production:
v2.0.0, Pocket ID hairpin, OIDC client from vault_starpunk.pocketid and
the admin sub in starpunk_admin_sub.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Karakeep had DISABLE_SIGNUPS unset, so anyone could make an account on
bookmarks. and use the OpenAI key behind it. The server also had no
automatic updates and three security updates waiting; the common role now
installs unattended-upgrades with Debian's default security-only origins.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The READMEs, CLAUDE.md and todo.md still described Authentik, services
removed long ago (Ghost, MMDL, Postiz, Pinchflat...) and two servers. Now
they list the 24 services actually deployed (checked against docker ps),
explain that netcup and docker-01 are one machine, document Gondulf,
StarPunk, Listmonk, Gokapi and dockcheck, and use bin/deploy throughout.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
--deployment-password sets the password and exits, so the detached run
plus health wait never saw a healthy container.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The 'Gokapi started' log line is not always flushed to log.txt, so the
first-run wait timed out even though the super admin had been created.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
All of Authentik's applications (including Tailscale's tailnet login)
have moved to Pocket ID, and its only recent sign-ins were those
handovers. Drops its stack, the auth. route and DNS record, the auth
subdomain variables, Gotify's leftover hairpin entries, and the dockcheck
exclude; the docs now name Pocket ID as the SSO provider.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Gokapi now keeps its config and data in /opt/stacks/gokapi instead of named
volumes. config.json is templated with Pocket ID as the OIDC provider, so a
fresh install never shows the open /setup wizard. On a new database a one-off
run with --deployment-password creates the super admin, after which only
existing users can log in.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>