Merge pull request 'Add CLAUDE.md, collection scaffold, and verified odin access docs' (#1) from docs/repo-scaffold-and-claude-md into main

Reviewed-on: #1
This commit was merged in pull request #1.
This commit is contained in:
2026-08-15 20:34:24 -04:00
8 changed files with 520 additions and 0 deletions
+11
View File
@@ -0,0 +1,11 @@
# Secrets — every collection keeps its own .env beside its .env.example
.env
**/.env
!**/.env.example
# MemPalace per-project files (issue #185).
# `mempalace init` ignores both by default. We deliberately track mempalace.yaml —
# it is just wing/room config, and losing it on a fresh clone would make mining
# fall back to basename auto-detection (wing "scripts", which collides).
# entities.json stays ignored: it holds detected people/entities.
entities.json
+212
View File
@@ -0,0 +1,212 @@
# scripts
Operational scripts for Luc's self-hosted environment. Each script collection is an
independent, self-documenting folder at the repo root. There is no shared runtime, no
build step, and no top-level package manifest — this is a repo of standalone tools, not
an application.
---
## The one hard rule: collections are self-contained
Every script collection lives in **its own top-level folder** and carries **all of its own
documentation, configuration, and dependencies**. A collection must be understandable,
runnable, and deletable on its own.
```
scripts/
├── CLAUDE.md # this file — repo-wide context
├── README.md # human-facing index of collections
├── _template/ # scaffold for a new collection (leading _ = not a collection)
└── <collection>/
├── README.md # REQUIRED — what it does, prerequisites, usage
├── CLAUDE.md # OPTIONAL — agent context, only if non-obvious
├── .env.example # REQUIRED if the collection reads secrets
├── bin/ # executable entrypoints (chmod +x, shebang)
├── lib/ # sourced helpers, not directly executable
└── docs/ # anything longer than the README warrants
```
**Do not** create shared/, common/, or utils/ at the repo root. If two collections need
the same helper, copy it. Duplication across collections is explicitly preferred over
coupling — a collection that depends on a sibling is no longer independently deletable,
and these scripts get deployed to different places at different times.
Folders prefixed with `_` are repo infrastructure, not collections.
### Adding a collection
Copy `_template/` to the new name, fill in its README, and add one line to the root
`README.md` index. That is the whole process.
---
## The environment
**odin** — the Unraid server; the single host everything self-hosted runs on. Reached from
this box over **Tailscale** (`100.101.253.105`, LAN `192.168.90.103`).
> **See [docs/odin-access.md](docs/odin-access.md)** for the access path, the WebGUI/API
> endpoint, TLS caveats, and the Unraid GraphQL API — all verified from this box.
Verified characteristics:
| Aspect | Detail |
| --- | --- |
| Host OS | Unraid |
| Container runtime | Docker + Docker Compose (stacks defined per-app) |
| Database | MariaDB 11.x — **PostgreSQL is not available**, treat as a hard constraint |
| Cache / queue | Redis (available, used optionally) |
| Auth | Authelia — OIDC/OAuth2 provider for internal apps |
| Ingress | Pangolin/Newt tunnel — **no open inbound ports** |
| DNS | Split-DNS on `bergerhouse.net`; private IPs resolve internally |
| Git forge | Self-hosted Gitea/Forgejo at `git.bergerhouse.net` (user `luckberg`, SSH port 222) |
| CI | Gitea Actions with a self-hosted runner |
| Mail / calendar | Fastmail (paid) — JMAP/CalDAV, the source of truth for calendars |
| Remote access | Tailscale; odin `100.101.253.105`, subnet route for `192.168.90.0/24` |
| Unraid API | GraphQL at `{myunraid.net host}:8443/graphql`, `x-api-key` auth (Unraid 7.2+) |
| Automation | n8n on odin — drives the Docker-manager Telegram bot |
Known service hostnames: `git.bergerhouse.net`, `familysync.bergerhouse.net`,
`familysync-dev.bergerhouse.net`.
**Not on odin:** GitHub (`gh` is not installed — use `tea` for the Gitea forge),
PostgreSQL, any cloud provider.
### This box vs. odin
The machine this repo is edited on is `dev` (Tailscale `100.94.16.46`) — a **separate
workstation**, and in fact a VPS with a public IP, not a machine on the home LAN. Scripts
written here usually *target* odin. Never assume a script runs on the same host it was
authored on: take the target as a parameter or read it from config.
Tailscale is the **only** path from `dev` to odin — if it is down, nothing below resolves.
`?` — still unconfirmed: the SSH login to odin (no `~/.ssh/config` entry on this box) and
where deployed scripts are expected to live on the Unraid host.
### Sibling repos on `git.bergerhouse.net`
`luckberg/` holds `scripts` (this repo), `familysync`, `unraid-docker-manager`,
`azure-epac`, `_cargo-index`. Two are load-bearing context:
- **`unraid-docker-manager`** — n8n + Telegram bot managing odin's containers. The
authoritative source on the Unraid GraphQL API and container polling; distilled into
[docs/odin-access.md](docs/odin-access.md).
- **`familysync`** — the origin of most of the stack facts in the table above.
---
## Prior context: MemPalace
Much of what is known about this environment lives in **MemPalace**, a local semantic
memory store — not in this repo. Query it before asking Luc to re-explain something, and
before assuming an infrastructure detail is undocumented.
| Aspect | Detail |
| --- | --- |
| Version | `mempalace` v3.4.1, installed as a `uv` tool |
| Palace path | `/home/luc/.mempalace/palace` |
| Backend | ChromaDB (vector search) + SQLite (metadata); fully local, no API key |
| Access | MCP server (`mempalace-mcp`) → 19 `mempalace_*` tools; plus the `mempalace` CLI |
| Size | ~32k drawers |
**Read the docs, don't guess the CLI.** MemPalace ships its own instructions:
```bash
mempalace instructions <help|init|mine|search|status>
```
Structure is **Wings → Rooms → Closets → Drawers** — wings are projects/people, rooms are
topics, closets are summaries, drawers are verbatim memories. *Halls* connect rooms within
a wing; *tunnels* connect rooms across wings.
Existing wings: `familysync` (~21.6k), `sessions` (~10.6k — mined agent transcripts),
`wing_familysync`.
### This repo's wing
Initialized with `mempalace init . --yes` and mined with `mempalace mine .`.
- **Wing: `odin-scripts`** — set explicitly in `mempalace.yaml`, *not* auto-detected.
Auto-detection uses the directory basename, and a wing called `scripts` would collide
with any other corpus in a folder of that name.
- **Rooms:** `environment`, `documentation`, `collections`, `general`.
- Room routing matches keywords against the file **path**, so never use `scripts` as a
room keyword here — it matches this repo's own path and swallows every file. (Learned
the hard way; the comment in `mempalace.yaml` says so too.)
- `mempalace.yaml` and `entities.json` are gitignored — that is MemPalace's own
convention (issue #185), applied by `init`.
Re-mine after adding a collection: `mempalace mine . --agent claude`. Mining is
idempotent — already-filed files are skipped.
Practical notes:
- `mempalace_search` is semantic — the `query` field takes **keywords only**, max 250
chars. Put background in `context`, not in the query, or the embedding gets diluted.
- Filter with `wing`/`room` when you know where the answer lives; the `sessions` wing is
raw transcript chunks and is noisy for factual lookups.
- **BM25 scores of 0 with mid-range similarity means no lexical match** — a bare-keyword
search for a proper noun that returns only semantic neighbours is a *negative* result,
not a weak positive. This is how the `odin` gap above was established.
- `mempalace_kg_query` hits the temporal knowledge graph for point-in-time facts; prefer
it over `search` for "what is X currently" questions, and use `kg_supersede` rather
than invalidate-then-add when a single-valued fact changes.
- Auto-save hooks are active: a **Stop** hook saves every 15 human messages, and a
**PreCompact** hook force-saves before context compaction.
Complementary and **separate** from MemPalace:
- **Per-project agent memory** — `~/.claude/projects/<slug>/memory/*.md`, indexed by that
directory's `MEMORY.md`. This repo's is currently empty; FamilySync's holds the CI-gate,
dev-stack, and Gitea-CI notes that much of this file's environment section derives from.
- **Project docs on disk** — `~/projects/familysync/CLAUDE.md` and `docs/ARCHITECTURE.md`
are the authoritative written source for the odin stack.
---
## Conventions
These apply to every collection unless its own README documents a deliberate exception.
### Shell
- `#!/usr/bin/env bash` and `set -euo pipefail` at the top of every bash entrypoint.
- Quote every expansion. Prefer `[[ ]]` over `[ ]`.
- Scripts must be **idempotent** — safe to re-run. Assume cron or a retry will do so.
- Accept `--dry-run` for anything that mutates state, deletes, or sends. Default to the
safe path when a flag is ambiguous.
- Log to stdout, errors to stderr. No log files unless the collection documents rotation.
### Secrets
- **Never commit secrets.** Read them from the environment or a file path passed in.
- Every collection that needs secrets ships a `.env.example` with dummy values and a
README line naming where the real values live.
- Note the FamilySync-side precedent: gitleaks runs as a blocking CI gate on that repo.
Assume any secret committed here is treated the same way — as a leak to be rotated, not
a mistake to be amended away.
### Portability
- Target the interpreters actually present: bash, Python 3, Node 22. Check before adding
a dependency on anything else.
- Unraid's userland is BusyBox-leaning in places — prefer POSIX-portable invocations of
`sed`/`awk`/`date` over GNU-only flags when the script runs on the server.
---
## Working agreements for Claude
- **Scope changes to one collection.** A request about collection X should not touch
collection Y. If it must, say so explicitly rather than doing it quietly.
- **Update the collection's README in the same change as the code.** Docs living beside
the script is the point of this repo's layout; a code change that leaves its README
stale defeats it.
- **Don't invent infrastructure.** If a script needs a host, path, port, or credential
that isn't documented above, ask — do not guess a plausible value. Wrong infra
assumptions in an ops script fail in production, not in review.
- **Destructive operations need confirmation** before they're run against odin, even when
the script itself is finished and correct. Writing the script is not authorization to
execute it.
+26
View File
@@ -0,0 +1,26 @@
# scripts
Operational scripts for my self-hosted environment (**odin**, an Unraid host running
Docker Compose behind Authelia and a Pangolin/Newt tunnel).
Each script collection lives in its own top-level folder and is fully self-contained —
its own README, config, and dependencies. Collections never import from each other.
See [CLAUDE.md](CLAUDE.md) for the environment details and repo conventions, and
[docs/odin-access.md](docs/odin-access.md) for how to reach odin and drive its API.
## Collections
_None yet._
<!-- Add one line per collection:
| [name](name/) | What it does | Runs on |
-->
## Adding a collection
```bash
cp -r _template <collection-name>
```
Fill in its `README.md`, then add a line to the table above.
+3
View File
@@ -0,0 +1,3 @@
# Copy to .env and fill in. Never commit the filled-in .env.
EXAMPLE_HOST=odin.bergerhouse.net
EXAMPLE_TOKEN=replace-me
+36
View File
@@ -0,0 +1,36 @@
# <collection-name>
One sentence: what this collection does and why it exists.
## Prerequisites
- Where it runs: `odin` (Unraid) / this workstation / either
- Interpreters: bash / Python 3 / Node 22
- External tools: e.g. `docker`, `tea`, `curl`
- Access needed: e.g. SSH to odin, Gitea API token
## Configuration
Copy `.env.example` to `.env` and fill it in. Real values live in `<where>`.
| Variable | Required | Description |
| --- | --- | --- |
| `EXAMPLE_HOST` | yes | Target host to operate against |
## Usage
```bash
bin/example --dry-run # preview, changes nothing
bin/example # apply
```
## Behaviour notes
- Idempotent: yes/no — and what happens on a re-run.
- Destructive operations: list them, or state "none".
- Scheduling: cron entry / Unraid User Scripts / manual only.
## Gotchas
Anything that cost time to discover. This is the section that earns the collection its
own folder — keep it honest and current.
+55
View File
@@ -0,0 +1,55 @@
#!/usr/bin/env bash
# Template entrypoint. Copy, rename, and replace main().
set -euo pipefail
readonly SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
readonly COLLECTION_DIR="$(dirname "$SCRIPT_DIR")"
DRY_RUN=0
usage() {
cat <<'EOF'
Usage: example [--dry-run] [--help]
--dry-run Show what would happen without changing anything.
--help Show this message.
EOF
}
log() { printf '%s\n' "$*"; }
warn() { printf '%s\n' "$*" >&2; }
die() { warn "error: $*"; exit 1; }
# Run a mutating command, or describe it under --dry-run.
run() {
if (( DRY_RUN )); then
log "[dry-run] $*"
else
"$@"
fi
}
load_env() {
local env_file="$COLLECTION_DIR/.env"
[[ -f "$env_file" ]] || return 0
set -a; . "$env_file"; set +a
}
main() {
while (( $# )); do
case "$1" in
--dry-run) DRY_RUN=1 ;;
--help|-h) usage; exit 0 ;;
*) die "unknown argument: $1" ;;
esac
shift
done
load_env
: "${EXAMPLE_HOST:?EXAMPLE_HOST is not set — see .env.example}"
log "target: $EXAMPLE_HOST"
run true # replace with the real work
}
main "$@"
+149
View File
@@ -0,0 +1,149 @@
# Accessing odin
Reference for reaching the Unraid host (**odin**) from this workstation (`dev`) and
driving its API. Everything here was verified from `dev` on 2026-08-15; the "Verified"
column says how.
Related repo: **`luckberg/unraid-docker-manager`** on `git.bergerhouse.net` — an n8n
Telegram bot that manages odin's containers. It is the origin of most of the API detail
below and is worth reading before writing anything new against the Unraid API.
---
## Network path: Tailscale
`dev` and odin are both on the tailnet. `dev` is a **VPS** (public IP `51.222.234.162`),
*not* a machine on the home LAN — Tailscale is the only path to odin.
| Node | Tailscale IP | Notes |
| --- | --- | --- |
| `dev` | `100.94.16.46` | this workstation (linux) |
| `odin` | `100.101.253.105` | the Unraid host (linux) |
Other tailnet nodes seen: `duplicati`, `experience` (windows), `fedora`, `nusgwvm`,
`salt`, `google-pixel-8-pro`.
**A subnet route is advertised for the home LAN.** odin's LAN address `192.168.90.103` is
reachable from `dev` and routes over `tailscale0`:
```
$ ip route get 192.168.90.103
192.168.90.103 dev tailscale0 table 52 src 100.94.16.46
```
Both the tailnet IP and the LAN IP answer at ~34 ms RTT — the same path. This matters:
the `myunraid.net` URL below looks like a cloud relay, but **DNS resolves it to the
private `192.168.90.103`**, so traffic goes direct over Tailscale and never leaves the
tailnet. If Tailscale is down, none of it is reachable.
---
## WebGUI / API endpoint
```
https://192-168-90-103.87c90a69c53e7197560c778a63f483e675130f2a.myunraid.net:8443/
```
The hostname encodes odin's LAN IP with dashes, under a per-server hash on `myunraid.net`.
This is the **canonical vhost** — not merely a convenience alias. Use it for everything.
### Why the plain LAN IP does not work
Both failure modes are real and confirmed:
| Attempt | Result | Verified |
| --- | --- | --- |
| `https://192.168.90.103:8443/graphql` | **404** from nginx | curl |
| `http://192.168.90.103/graphql` | **302** → the `myunraid.net` URL | curl |
| `https://<myunraid-host>:8443/graphql` | **200**, GraphQL responds | curl |
nginx only serves `/graphql` on the `myunraid.net` `server_name`; the raw IP vhost returns
404. Plain HTTP redirects to that hostname — and **the redirect strips the `x-api-key`
header**, so a client that follows redirects authenticates as nobody. Always call the
`myunraid.net` URL directly rather than relying on a redirect.
### TLS caveat — correcting the other repo
`unraid-docker-manager` states "Valid certs via myunraid.net — no SSL ignore needed."
**That is not true from `dev`.** Verification fails here:
```
depth=1 C=US, O=Let's Encrypt, CN=YR1
verify error:num=20:unable to get local issuer certificate
depth=0 CN=*.87c90a69c53e7197560c778a63f483e675130f2a.myunraid.net
```
The leaf is a genuine Let's Encrypt wildcard, but the chain presents an intermediate
(`YR1`) that this box's CA bundle (`ca-certificates 20250419`) cannot chain to a trusted
root — either the server omits the intermediate or the bundle predates it. The cert is not
forged; the chain is incomplete *for this client*.
Practical consequence: **`curl` needs `-k` from `dev`**, and any script that talks to odin
must either pass the insecure flag or pin/supply the issuer. n8n running *on* odin may
well verify fine, which is likely why the other repo's claim held there. Prefer supplying
the missing intermediate over making `-k` a permanent habit — `-k` disables verification
entirely, which on a tailnet is a tolerable but real weakening.
---
## Unraid GraphQL API
- **Endpoint:** `{UNRAID_HOST}/graphql`, POST, where `UNRAID_HOST` is the `myunraid.net`
base URL **without** the `/graphql` suffix.
- **Auth:** `x-api-key: <key>` header.
- **Availability:** native in Unraid 7.2+ (odin is on 7.2); 6.97.1 needs the Connect plugin.
Unauthenticated requests return HTTP **200** with a GraphQL error body — not an HTTP 401:
```json
{"errors":[{"message":"Invalid CSRF token","extensions":{"code":"UNAUTHENTICATED",
"originalError":{"error":"Unauthorized","statusCode":401}}}],"data":null}
```
**Never treat HTTP 200 as success.** Always inspect `response.errors[]`.
### Polling running containers
```bash
curl -sS -k -X POST "${UNRAID_HOST}/graphql" \
-H 'Content-Type: application/json' \
-H "x-api-key: ${UNRAID_API_KEY}" \
-d '{"query":"query { docker { containers { id names state } } }"}'
```
```json
{"data":{"docker":{"containers":[
{"id":"<server_hash>:<container_hash>","names":["/n8n"],"state":"RUNNING"}
]}}}
```
Field behaviours that bite:
- `state` is **UPPERCASE**`RUNNING`, not `running`.
- `names` is an array and entries are **`/`-prefixed** — `/n8n`, not `n8n`.
- `id` is a `PrefixedID`: `{server_hash}:{container_hash}`, two 64-char SHA-256 hex
strings joined by a colon. The server half is identical for every container on odin.
- **`isUpdateAvailable` does not exist** in the 7.2 schema, despite appearing in that
repo's earlier research. Introspect before relying on any field.
### API keys
Not present on this box — `.env.unraid-api` is gitignored and absent from the clone. Create
a key scoped to what the script actually needs:
```bash
unraid-api apikey --create --name "<name>" --permissions "DOCKER:UPDATE_ANY" --json
```
Or WebGUI → Settings → Management Access → API Keys. Store it in the collection's own
`.env`, never in git.
---
## Gotchas carried over from `unraid-docker-manager`
- **Each Bash tool call is a fresh shell.** Source the env file in the *same* command
chain as the request: `. .env.unraid-api; curl ...`. A separate `source` call is lost.
- **Large API responses:** save to a temp file before parsing; piping a 400 KB body
straight into `python3 -c` fails silently.
- Docker API `204 No Content` means success with an empty body.
+28
View File
@@ -0,0 +1,28 @@
# Wing is set explicitly, not auto-detected from the directory basename —
# "scripts" is a common folder name and would collide with other corpora.
wing: odin-scripts
rooms:
- name: environment
description: Repo-wide context — the odin/Unraid environment, conventions, agent guidance
keywords:
- environment
- odin
- unraid
- conventions
- CLAUDE
- name: documentation
description: Per-collection READMEs and docs/ content
keywords:
- documentation
- docs
- readme
# NB: do not use "scripts" as a keyword here — it matches the repo's own path
# (/home/luc/projects/scripts) and would swallow every file into this room.
- name: collections
description: Executable entrypoints and sourced helpers (bin/, lib/)
keywords:
- bin
- lib
- name: general
description: Files that don't fit other rooms
keywords: []