rundesk-ai/rundesk-cli / src/skills/managing-rundesk
managing-rundesk
VerifiedUse when managing this agent, any other agent in its Rundesk install, or the install itself. It supplies verified operational workflows and safety rules for agent instructions, gateways, conversations, channels, schedules, providers, skills, backups, configuration, credentials, and the Rundesk lifecycle. Do not use for ordinary project work or for deciding, performing, or reviewing delegation.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 5.0 KB |
| references/agent-instructions.md | markdown | 12.3 KB |
| references/agents.md | markdown | 5.9 KB |
| references/backups.md | markdown | 2.1 KB |
| references/channels.md | markdown | 6.6 KB |
| references/configuration.md | markdown | 3.6 KB |
| references/conversations.md | markdown | 5.3 KB |
| references/gateways.md | markdown | 3.1 KB |
| references/lifecycle.md | markdown | 3.7 KB |
| references/maintenance.md | markdown | 5.9 KB |
| references/oauth-login.md | markdown | 3.4 KB |
| references/permissions.md | markdown | 3.6 KB |
| references/providers.md | markdown | 2.8 KB |
| references/retrospective.md | markdown | 3.8 KB |
| references/schedules.md | markdown | 6.4 KB |
| references/self-improvement.md | markdown | 10.7 KB |
| references/skills.md | markdown | 6.6 KB |
| references/teams.md | markdown | 3.2 KB |
references/oauth-login.md
View on GitHub# OAuth login
Use `"$RUNDESK_COMMAND" login --help` as authoritative. Provider-specific app creation, API
enablement, consent-screen scopes, and resource permissions live in the installed provider skill,
never here.
## The ordinary path
```sh
"$RUNDESK_COMMAND" env set <PROVIDER>_OAUTH_CLIENT_ID
"$RUNDESK_COMMAND" env set <PROVIDER>_OAUTH_CLIENT_SECRET
"$RUNDESK_COMMAND" login <provider>
```
The app client is the owner's value, placed with `env set` like any other and typed rather than
passed as an argument. `<PROVIDER>` is the declared provider ID as a shell variable, so `google`
gives `GOOGLE_OAUTH_CLIENT_ID`. Placing it first is normal, including before the provider's catalog
is installed; `login` then asks for nothing. It prompts only for a value that is genuinely missing.
Repeat `login` to connect more accounts to the same app. An account is filed under its immutable
subject, so a changed address is the same connection. Integrations select the account with
`--email`; a missing declared capability scope may reopen consent for that same account.
Never ask anyone for a refresh token, and never put a client secret in a skill, in argv, or in
output.
## Profiles are for a second app, and are rarely needed
`--profile <name>` selects a different OAuth *app client*, not a different account, and suffixes the
same names (`GOOGLE_OAUTH_CLIENT_ID__WORK`). One app with several connected accounts is the normal
shape — use `--email` to choose between them. Reach for `--profile` only when there are genuinely
two apps.
## Replacing an app client
```sh
"$RUNDESK_COMMAND" login <provider> --replace-client
"$RUNDESK_COMMAND" login <provider> --replace-client --confirm
```
Without `--confirm` this only previews: it counts the grants that would be discarded, prompts for
nothing, and changes nothing. With it, the new client signs in first and only then replaces the old
client and its grants together — a replacement that cannot sign in leaves everything as it was.
Only the named profile is touched. `--confirm` without `--replace-client` is refused.
## What the browser does
`login` prints the exact loopback address it is listening on, then opens a browser. The callback is
`http://127.0.0.1:<ephemeral-port>/<random-path>`: the literal address rather than `localhost`, a
port chosen per sign-in, and a path and state checked exactly. Stray local requests are answered and
ignored until the real redirect arrives or the deadline passes. The success page tells the person to
return to the terminal.
With no browser to open, the authorization URL is printed to follow by hand. It carries that flow's
short-lived state and PKCE challenge — that is the request — and never a client secret,
authorization code, refresh token, or access token. Manual copy-and-paste of a *code* is not
supported.
## When it refuses
Retry declined, timed-out, wrong-account, and concurrent-consent failures; none of them writes a
grant. A missing client is refused by name, saying exactly which values to set. A changed
declaration is refused until reviewed and reconnected. A revoked grant is reported as such and names
the login command that fixes it. One malformed installed declaration disables only its own provider.
Grants are sealed under a name no `rundesk env` verb reads or writes. App client values are ordinary
owner values, visible in `env list` as a hint, but withheld from what a turn is handed. Backups
include the local key beside them and need protection.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |