What you are approving
Expires
Loading…
Requests from Claude
Claude can only ask. Nothing below is authorized or changed until you approve it.
The last action may have completed. Refresh to check the saved state before another action. Pending authorizations can be resumed with their original request.
Configuration is read-only for this account.
Save a draft, then preview and apply it. Active channels keep their approved settings until their new configuration is authorized.
Loading configuration…
Saved settings changed. Your draft is retained. Refresh to inspect the current revision, then choose whether to keep your edits or use the saved draft.
Supported models
Browse providers before adding API keys. A model becomes selectable above when an eligible account is configured for this workspace.
Skill library
Reusable instructions. Selecting a skill does not grant tools or account access.
Import a repository skill set
Import every skill under a folder, including its supporting files, at one pinned commit. Scripts do not run automatically.
Reading repository…
Importing adds these revisions to your skill library. Select them for an agent before applying its configuration.
No matching skills. Create instructions that your agents can reuse.
Loading on demand requires a supported conversation tool runtime. Preview checks availability.
Conversation memory
These notes belong to their original channel, thread, participant and authorization. They are not shared workspace knowledge. Each scope holds up to 32 notes and 16,000 bytes.
Loading saved notes and proposals…
No saved notes or proposals for this agent.
Showing the first 100 entries and proposals.
Pending review
No proposals awaiting review.
Saved notes
Research
Research topics using Hacker News discussions, Polymarket markets, and YouTube titles and metadata.
Choose the research sources and page reading this agent can use. Open web and full page reading require a Claude SDK model. Publishing changes requires channel reauthorization.
Installation starts disabled. Enable it after reviewing its sources, then select it and apply the configuration. General code and MCP plugins require a supported runtime.
Tuneplane Studio
Send TikTok links to the curation inbox, organize titles and tags, and approve items for the off-air library.
Requires a dedicated Tuneplane service connection for this agent. Approval prepares the library; publication stays in Studio. Publishing configuration changes requires channel reauthorization.
Anyplane Delivery
Report what is deployed and in flight across Anyplane repositories: deployed versions, default-branch CI, running workflows, open pull requests and pending releases.
Read-only. Requires a dedicated Anyplane Control Center service connection for this agent, and cannot merge, deploy or change anything. Publishing configuration changes requires channel reauthorization.
Anyplane Release
Lets a merge agent merge pull requests and a deploy agent deploy and roll back Anyplane applications, each through the release workflows the repository already has.
These agents can change production. They act only on requests from an owner or admin, only on applications in the reviewed target list, and nothing runs until you switch release effects on under Release requests. A rollback also needs your approval there. Credentials and the release service are configured separately by your operator. Publishing configuration changes requires channel reauthorization.
Revision history
Effective configuration preview
Unsaved changes
Review affected agents
Applying publishes these agent settings. Each listed Slack channel keeps its approved configuration until its successor is authorized.
Published
The last change may have completed. Refresh to check the saved state before another change.
Release effects
Loading…
While off, requests are recorded and checked but no agent merges, deploys or rolls back anything.
The merge and deploy agents can act on requests from owners and admins. Turning this off stops the next action immediately; work already done stays recorded.
Only an organization owner can turn release effects on.
Turn on release effects?
The merge and deploy agents will merge pull requests and deploy applications for requests from organization owners and admins, within the reviewed target list. A rollback still needs your approval here. You can turn this off at any time.
Rollbacks waiting for approval
Requests
Requests recorded for the merge and deploy agents, newest first. Text written by requesters and agents is shown as-is and is never an instruction.
Loading…
No release requests yet.
| Request | Status | Agent | Requested by | Created | Open |
|---|---|---|---|---|---|
Outcome:
Written by the agent. Untrusted text:
Note from the requester. Untrusted text:
Reason from the requester. Untrusted text:
Message from the requester. Untrusted text:
Rollback approval
These are the facts the deploy agent recorded from its own reads. Your decision is bound to exactly these facts.
Facts digest:
The recorded facts do not match their digest. This approval cannot be decided here.
Part of the requester's reason is too long to show in full, so this rollback cannot be approved here. You can still deny it.
This dashboard has no approval signing key, so a rollback cannot be approved here. You can still deny it.
Only you can make this statement. Neither the requester nor an agent can. An approval lasts 30 minutes after you give it and can be used once.
Only an organization owner can decide a rollback.
What was done
| Action | Result | References | When |
|---|---|---|---|
| Workflow run |
An action marked indeterminate may or may not have happened; it is never repeated and is checked by later reads.
Waiting for
Messages
Untrusted text.
Timeline
The last change may have completed. Refresh to check the saved state before another change.
You do not have access to code review in this organization. Ask an organization owner if you need it.
Organization policy
The default every repository in this workspace inherits. A repository can add rules, never remove them.
Only an organization owner can change the policy.
This is the current policy. There is nothing to save.
This loosens the organization default. Repositories in this workspace will be reviewed less strictly: .
Repositories
Repositories whose review runs speak for this workspace. Each is pinned to the review workflow versions listed.
Activation code, shown once. Store it as an environment secret for the repository’s activation workflow. It cannot be shown again.
No repositories are registered for this workspace yet.
Pending repositories cannot be activated from this dashboard yet. Owner activation is limited to first-party organizations until repository ownership can be proved through GitHub.
| Repository | Status | Allowed workflow refs | Registered | Actions |
|---|---|---|---|---|
| Activated by an ownerActivated by its workflow | |
Review activity
Computed from the metadata each review run uploads. No code or comments are stored.
Loading review activity…
Only the newest uploads in this window are included, so older rounds may be missing from these numbers. Choose a shorter window for exact figures.
No review runs uploaded results in this window.
Gate conclusions by tier
Each pull request counts once, by its latest round.
| Tier | Passed | Failed | Neutral | Pending | Total |
|---|---|---|---|---|---|
Rounds per pull request
| Rounds | Pull requests | Share |
|---|---|---|
Findings and their outcomes
| Bug class | Findings | Noise | Fixed |
|---|---|---|---|
|
Circuit breaker
State not evaluated here
Whether it is raising the tier depends on merged pull requests and their reverts, which review uploads do not carry. The review Action evaluates it nightly and reports it as the review/circuit-breaker check on each repository’s base branch.
Recent uploads
| Pull request | Round | Tier | Gate | Findings | Received |
|---|---|---|---|---|---|
|
Solo-owner exception |
Health
All checks
Your agent runs on the verified key. It has no tools or plugins.
Slack is connected and your agent can see its channel.
Before you leave for Slack
- Sign in to Slack as an owner or admin of the Slack workspace you want. A request from a regular member cannot finish this setup.
- The channel list can be empty. Slack lists only channels the Missionplane app has joined. After the install, add the app to a channel (type /invite in the channel and pick the app), then reload the list.
- Do not remove the Missionplane app from Slack afterwards. Removing it stops the agent, and you have to connect again.
In your channel, send:
This page now checks every 30 seconds.
Credits and usage
Loading usage…
Usage could not be loaded. Missionplane did not get a usable answer. Nothing here is a balance of zero. Try again in a moment.
You cannot see usage for this workspace. Your role does not include usage. Ask an owner or admin of this organization.
Usage is only shown to the person who created this workspace. This workspace is not available to you here, or it no longer exists.
Balance
credits available
Your balance is zero. New runs funded by credits are refused until you add credits or switch to your own model key.
Free allowance
The allowance is close to its end. Unused allowance does not carry over.
Balance unavailable
Missionplane could not read your credit balance just now. This does not mean your balance is zero, so no number is shown. It is read again every minute; check back shortly.
Setting up credits
Missionplane is preparing your credit account. The balance appears here when it is ready, usually within a few minutes. This page checks by itself.
Credits could not be set up
Your own model key still works. Contact Anyplane support and quote this reference:
Credits are not available for this organization
Your runs use your own model key. Contact Anyplane support if you expected credits.
No credit balance
Your runs use your own model key, so your provider bills the model use and there are no credits to count. The credit figures below are an estimate of what the same runs would cost in credits.
Credit spending is switched off for this workspace. New runs funded by credits are refused until it is switched back on.
Pay for runs with
Anyplane credits: available when billing opens.
Limits
- Daily limit
- Per-run limit
- Also applies to your own key
- New runs
Confirm these limits
| Setting | New | Now |
|---|---|---|
Usage in this period
No runs in this period yet. Usage appears here after your agents run.
- Model tokens
- Credit equivalent
- Charged to credits
- Turns not priced
The credit equivalent is what the priced turns would cost in credits. Only runs paid with Anyplane credits are charged; runs on your own key are billed by your provider.
| Paid with | Runs | Tokens | Credit equivalent | Charge |
|---|---|---|---|---|
By agent
Covers the newest runs listed below, not the whole period.
Loading agents…
| Agent | Runs | Tokens | Credit equivalent | Charge |
|---|---|---|---|---|
Runs
On small screens, scroll the table horizontally for the other columns.
| Run | Outcome | Paid with | Model | Tokens | Credit equivalent | Charge |
|---|---|---|---|---|---|---|
|
Day by day (UTC)
| Day | Runs | Tokens | Credit equivalent | Charged to credits |
|---|---|---|---|---|
Apps for agents
Loading apps…
Apps are not available for this workspace.
Your role does not include agent configuration for this workspace, so apps cannot be shown.
Apps could not be loaded. Nothing was changed.
Save or discard your choices before choosing another agent.
This workspace has no agents yet. Create one from Agents, then come back to give it apps.
Loading this agent’s apps…
Only a workspace administrator with agent configuration access can change an agent’s apps. You can see what is set.
Saved to the agent’s draft. The agent does not use these apps yet. Open its settings to preview and apply them. Applying changes what the agent may do, so Slack asks you to authorize it again.
No apps are offered yet.
This app is no longer offered. Turn it off and save to remove it.
What the agent can read
At least one is required. Up to 10, separated by commas. In Slack, open a person’s profile, choose the three dots, then Copy member ID. A change is approved by the person who asked for it in Slack, and only if they are listed here. If someone who is not listed asks for a change, the agent is refused.
How long an approval card stays open. 1 to 15 minutes; 10 by default.
Review before saving
Saving writes the agent’s draft only. The agent keeps using what it has until you apply the draft in its settings. Applying changes what the agent may do, so Slack asks you to authorize it again before it can use the new tools.
The last change may have completed. Refresh to check the saved state before another change.
Agent writes
Loading…
While paused, no agent changes anything. Resuming needs an owner to approve a resume request on this page, with an MFA check.
Agents may act. Pausing is immediate, needs no MFA check, and is open to owners and admins.
Waiting for a decision
Critical actions outside a pull request. The facts come from the system that asked, never from an agent. Approving or denying asks for an MFA check first.
Loading…
Nothing is waiting for a decision.
Approve all pending batches
One MFA check approves every pending pull request batch below, each with the pull requests still ticked in its list. Each batch is still approved and signed on its own, for 24 hours, at the head commits shown. Untick a pull request in its batch to leave it out; a batch with nothing ticked stays pending.
Facts digest:
The recorded facts do not match their digest. This request cannot be decided here.
Each pull request is bound to the head commit shown. The titles, rules and links come from GitHub and review-kit; untick a pull request to leave it out of the approval (it stays on the GitHub review path).
| Include | Pull request | Head commit | Matched rules | Review |
|---|---|---|---|---|
|
review-kit findings |
This dashboard has no approval signing key, so nothing can be approved here. You can still deny.
Only an organization owner can decide.
Recent decisions
Approved, denied and expired requests, newest first.
| Request | Status | Decided by | Use | Requested |
|---|---|---|---|---|
Not enabled for your organization
Workspaces, agent configuration, conversations, and durable run history are available from the navigation.
Agent conversations in Slack
Choose an agent, connect with Slack, then select channels by name.
Your organization and Missionplane workspace are filled in automatically. Slack access and agent authorization are reviewed separately.
Checking Slack setup…
No editable agents in this workspace. Open Agents to create one or review your access.
Slack conversations are unavailable for this agent’s current configuration.
Slack setup is unavailable. Contact your workspace administrator.
Models for Slack conversations
Enable the models your agents can use in this workspace’s Slack connections.
Your operator configures model connections and stores provider credentials securely. Workspace administrators choose which of those models are enabled for each Slack connection.
These settings do not change an agent’s saved model or its Slack authorization. Disabling a model blocks new turns and model calls using it. Work already started may finish.
Loading Slack models…
No Slack connections in this workspace. Connect a workspace above, then refresh models.
Slack conversations
Channels where people can converse with an agent. Report delivery connections are listed separately below.
No conversational Slack connections in this workspace. This does not affect configured report delivery.
Selections show existing connections. They do not authorize a different channel or agent.
Remaining today
Authorization
This Slack connection is disabled. New runs cannot start.
Connection details match. Access to Slack and the agent is checked before each run.
Refresh to check the saved state before another change.
Authorize a Slack workspace with Relayplane, then return here and refresh. Conversational access and report-posting access are configured separately.
Open Relayplane setupAdditional plugins
General plugin installation is not enabled for your organization.
Research
A skill for agents that produce scheduled, cited reports.
Research topics using Hacker News, Polymarket, YouTube and the open web, then deliver a cited report or LinkedIn connection drafts to an authorized Slack channel.
Hacker News uses each topic’s lookback window. Polymarket returns active markets and YouTube results are not time-filtered. Open-web search runs on Anthropic’s side. Page reading opens only pages that research returned and never LinkedIn. Connection drafts are posted for you to review; nothing is sent. This skill does not enable general tools or interactive Claude chat.
Slack report delivery
Slack workspaces authorized for report posting. Choose a connection and channel in the agent configuration. These are distinct from your Missionplane workspace.
Connect with Slack, then choose channels by name. This opens a new tab and does not change any agent’s delivery destination.
Connecting alone does not change report delivery or enable a schedule.
Slack connection discovery is unavailable. Existing schedules can still be paused.
No Slack connections are available for this organization. Connect Slack, then refresh.
Model connections
Choose the default model for each connection. Credentials are managed securely outside the browser.
Default model:
Model access depends on the connected Claude account. A ready credential does not confirm access to every model.
No research model connection is configured for this organization.
Scheduled research
Scheduled research belongs to an agent. Its skill, connections, and authorization are configured from that agent.
Last attempt:
Run a report successfully before enabling this schedule.
Open the agent to resume the saved authorization request. Cancelling stops this automation.
No research automations in this workspace
Choose the Research skill when creating an agent to configure its schedule.
Choose a workspace
Open Workspaces to select a workspace and view its agents, conversations, and runs.
Archive ?
The server will check the current revision and active runs before archiving.
Organization workspaces
| Workspace | Status | Updated | Actions |
|---|---|---|---|
| Pilot managed |
No workspaces
No workspaces are available to this account.
This workspace
Records visible to you in this workspace.
Health
Your workspace
Claude Managed Agents
Connect an agent to a cloud or self-hosted environment, then run Claude sessions here.
Loading Claude configuration…
A previous request is awaiting confirmation. Its original request ID and inputs are retained. Refresh to check the current state before checking that same request again.
Review the message or interruption in the Claude Console. These requests cannot be recovered by entering a resource ID; keep the original request while its outcome is uncertain.
What will your agent do?
Start with a Missionplane template, write your own instructions, or connect an existing Claude agent.
Loading templates…
No matching templates. Try another search or start with your own instructions.
Your description becomes editable instructions in the next step.
Claude operations
Inspect environments and native schedules for this workspace.
Loading current Claude state…
Environment inventory requires agent configuration access. Your operator assigns cloud and self-hosted environments to this workspace.
No connections are assigned. Ask your Missionplane operator to enable Claude and assign a gateway connection for your organization.
No environments are assigned to this connection.
Cloud execution is managed by Claude. Worker polling metrics apply to self-hosted environments.
Checking worker readiness…
- Workers polling
- Pending
- Queue depth
This is a point-in-time observation. Refresh before starting work if worker availability has changed.
Your operator provisions and assigns environments. Worker enrollment and provider credentials stay with the operator; they are never entered into agent instructions.
Connect a Claude agent first to inspect deployments for its pinned version and environment.
Only deployments matching this connected agent's exact version and environment are shown. Native Claude schedules can run outside Missionplane's session controls.
Manual deployment · No recurring schedule
- Schedule
- Timezone
- Last run
The schedule is paused. These projected times do not mean a run will start.
No upcoming times reported.
Claude may start a scheduled run shortly after its listed time.
Workspace agents
| Agent | Skill and model | Status | Updated | Actions |
|---|---|---|---|---|
| Archived Pilot managed |
No agents
No agents in this workspace.
Claude environment and sessions
This agent executes in its Claude environment. Local executor configuration and standard conversations do not apply. Sessions have an approximate $5 provider budget; this workspace allows 10 new sessions and 100 managed mutations daily.
Claude Managed Agents are not enabled for this account.
Sessions
Create a session to start a task, or pick up a conversation.
You can review your sessions. Starting or messaging a session requires managed agent run permission.
Refreshing sessions…
Your first session starts here
Create a session above, then send Claude a task. Your conversations for this agent will appear here.
No sessions match this search and status.
Showing your latest recorded sessions for this connected agent. Open a session to refresh its status.
Opening session…
Session details
Claude session ID
Showing the latest 500 loaded events. Earlier events remain available in the Claude Console.
Instructions and model limits
Maximum response tokens: · Context:
This agent configuration is not available to edit here.
Slack conversations
Assign this agent to channels, then mention the Missionplane app to start a conversation. Each channel has one assigned agent, and replies stay in the thread.
Conversation access is separate from scheduled research reports.
Loading Slack settings…
Available channels could not be refreshed. Refresh Slack settings to try again.
This agent needs a supported model and registered tools before Slack conversations can be enabled.
Assigned channels
This agent has no assigned channels yet.
. Open that agent to resume the saved assignment. This channel's settings are locked until the assignment is complete.
The agent or Slack app permissions have changed. Reconnect Slack when prompted, then reauthorize this connection. Saved history stays available; after reconnecting, mention the agent again to start fresh conversation context.
Approve this agent's current instructions, model and tools for this channel. Conversation history and today's usage are kept.
Future new threads in this channel will use this agent instead of . Previous conversations remain in the old agent's history and are kept separate.
Approve this agent's current instructions, model and tools, with runs per day.
The agent has changed since this assignment was approved. The saved assignment cannot continue with different instructions.
Only an authorized connection owner can change these settings.
Agent overview
Last attempt:
Research skill and connections
Automation
Run a report successfully before enabling this schedule.
Research activity
No research runs recorded for this agent.
| Run | Trigger | State | Created | Delivery | Usage |
|---|---|---|---|---|---|
|
These attempts also appear in Runs. Delivery confirms posting; report bodies are not stored in this view.
Conversation history
Choose a conversation to inspect its stored messages.
No stored messages in this conversation.
Messages shown are durable history. A run that failed or has an uncertain outcome may not have a committed reply. Check Runs for its status.
Durable run progress; the completed conversation appears above after settlement.
The outcome is uncertain. This conversation cannot be replayed or continued until its outcome is resolved.
This tab retained only the request receipt. Message content was not saved in browser storage; check the original request to recover its outcome.
Execution history
Loaded execution history, newest first. Research includes the latest 100 visible attempts.
On small screens, scroll the table horizontally for delivery and inspection controls.
| Agent and run | Status | Trigger | Created | Delivery and usage | Actions |
|---|---|---|---|---|---|
|
No runs
No runs match this view.
Earlier research attempts exist. Open an agent to inspect its latest activity.
Research run
A failed or uncertain attempt is not automatically replayed. Review the agent’s current configuration before starting a new report.
Loading run details…
Run details
Replay is unavailable. Cancellation is offered only when the current run permits it.
The outcome is uncertain. Do not replay this run on the assumption that no effect occurred.
Loading the timeline…
What happened
Details expired. The step-by-step details of this run are no longer kept. The outcome above is still recorded.
Support reference
If you contact Anyplane support about this run, quote this reference. It identifies the run and contains no message content.
App tool calls
The app tool calls of this run could not be loaded. The run itself is unchanged.
Only the app, the operation and how each call ended are recorded here. Arguments and results are not kept, and an approval is shown by its state, never by who clicked.
-
Result code
Support references
Durable events
No visible durable events.