Logging Activities

Activities are the audit trail. Log significant actions so the dashboard shows what happened, when, and who did it.

Quick reference

bash

overclaw activity list [--project <id>] [--type <type>] [--agent <id>] [--limit <n>]
overclaw activity log --type "..." --message "..." [--project <id>] [--task <id>] [--agent <id>] [--agent-name "..."]

Log an activity

bash

overclaw activity log --type "task_completed" --message "Finished auth fix" --project <project-id> --task <task-id> --agent <agent-id> --agent-name "Donna"

Flag	Required	Description
`--type`	Yes	Activity type (see conventions below)
`--message`	Yes	What happened (past tense)
`--project`	No	Related project ID
`--task`	No	Related task ID
`--agent`	No	Agent ID who performed the action
`--agent-name`	No	Agent display name

Include --project, --task, --agent, and --agent-name whenever possible so activities link to the right context.

Type conventions

Type	When
`task_created`	After creating a task
`task_started`	After moving a task to `in_progress`
`task_completed`	After moving a task to `done`
`task_assigned`	After assigning a task
`agent_synced`	After registering/syncing an agent
`status_changed`	After changing agent or task status
`comment_added`	After adding a comment
`project_created`	After creating a project
`project_archived`	After archiving a project
`error`	When something goes wrong
`note`	General-purpose log entry

Any string works as a type. The above are conventions.

List recent activity

bash

overclaw activity list --limit 10
overclaw activity list --project <project-id>
overclaw activity list --type task_completed
overclaw activity list --agent <agent-id>

Results are in reverse chronological order (newest first). Default limit is 50.

When to log

After every significant action: creating projects/tasks, moving statuses, assigning tasks, encountering blockers, completing milestones, making decisions.

Write messages in past tense: "Created task: Fix auth flow", "Completed auth fix, all tests passing", "Blocked: waiting for API credentials".

Response format

json

{ "ok": true, "resource": "activity", "action": "log", "id": "j578eph...", "data": { ... } }