Agents are AI participants inside a Craken workspace. They talk with people, read the context you link for them, and use workspace tools to produce durable results such as wiki edits, summaries, file analysis, and follow-up messages. Every new workspace includes Orca; admins can add other server-managed agents, and external CLI agents connect through the CLI.
What Agents Can Do
Agents act through workspace tools:
- Read recent channel and DM context, search older messages, and answer in the current conversation.
- Find, read, write, revise, rename, restore, and verify Wiki pages when the right wiki workflow is loaded.
- Inspect Workspace files: read text and Markdown files, extract text from supported PDFs and office/data files, summarize the evidence, rename workspace files, and promote visible attachments into Workspace files when asked.
- Connect decisions across channels, wiki pages, workspace files, and messages.
- Search and inspect academic literature when the research skills are available.
Orca Capabilities covers the same ground for Orca, including the context boundaries it works inside.
Calling An Agent
Open Direct messages to Orca for private work, or mention @Orca — or another agent's name — in a channel where that agent is a member.
Agents stay out of ordinary human conversation unless they are mentioned or clearly addressed. If an agent does not answer in a channel, check the mention and the agent's channel membership.
Give Useful Context
A useful request names the goal, the source material, and the output you want. Link the source material with wiki, file, and channel link syntax instead of describing it.
Based on last week's discussion in #research, update the Risks section of [[Launch Plan]].
Use [[/briefs/q2-risks.pdf]] as evidence and return the result as a short table.
Recipes has worked examples spanning channels, wiki pages, files, and agents.
Jobs And Progress
Longer work becomes a Job: the conversation shows a progress row and a plan/checklist that the agent updates as the work proceeds. Send another message in the same conversation to change direction, or use the interrupt control on the progress row to stop. System operators can inspect additional Jobs diagnostics from Manage Workspace.
Token Limits
Agent usage is shared by the whole workspace and appears in the navigation footer. Simple questions use little of the quota; long, multi-step work uses much more. Near a limit, split the work into smaller requests or wait for the next daily or monthly reset.
Learning
Agents can learn stable workspace preferences and successful workflows over time. System operators can inspect Dreams diagnostics in Manage Workspace.