Follow these steps in order:

Step 1 — Read current state:

python3 ~/.duplocloud/bin/duplo_activate.py --check-state

Capture workspace and project from the JSON output. This makes no API call.

Step 2 — Workspace: resume or switch?

• If workspace is not null in the Step 1 output: ask the user:

  "You are currently working as <workspace.name>. Continue with this workspace? (y/n)"

  • y → skip to Step 5 (project resume check)
  • n → continue to Step 3

• If workspace is null: continue to Step 3.

Step 3 — List workspaces (also validates credentials):

python3 ~/.duplocloud/bin/duplo_activate.py --list-workspaces

• If exit code is 3 (no auth or invalid token): tell the user:

  "Your credentials file will open in VSCode. Fill in base_url and token, then save and close. If VSCode is not available, a terminal command will be shown instead."

  Then run:

  python3 ~/.duplocloud/bin/duplo_activate.py --edit-auth

  • If exit code is 0: re-run --list-workspaces.
    • If --list-workspaces still returns exit 3: tell the user "Credentials still invalid — please re-check the values." and re-run --edit-auth.
  • If exit code is 1 (VSCode not available): show the printed terminal command to the user and tell them to run it, then re-run /duplo:activate_project once they're done.

• If exit code is 0: show the numbered workspace list to the user. Write out each numbered item clearly in your response text so the user can read the full list without expanding tool output.

  • If the output contains "(Only one workspace found — will be auto-selected.)": automatically proceed to Step 4 with N=1 (skip asking the user).
  • If multiple workspaces are listed: ask the user:

    "Which workspace are you working as? (enter the number)"

Step 4 — Set workspace:

python3 ~/.duplocloud/bin/duplo_activate.py --set-workspace <N>

(Replace <N> with the number the user provided, or 1 if auto-selected.)

After setting the workspace, the old project is cleared from state. Continue directly to Step 6 (always show project list after an workspace change).

Step 5 — Project: resume or switch? (only reached if workspace was NOT changed in this run)

• If project is not null in the Step 1 output: ask the user:

  "Active project: <project.name>. Continue with this project? (y/n)"

  • y → show a summary of the active workspace and project, then continue to Step 5a.
  • n → continue to Step 6.

• If project is null: continue to Step 6.

Step 5a — Sync local files (same project):

Check if local files exist:

• .duplocloud/spec.md
• .duplocloud/plan.md

If neither file exists: silently download both (proceed to Step 5b).

If one or both exist: ask the user:

"Local spec/plan files exist. Overwrite with the latest versions from the platform? (y/n)"

• n → keep local files, skip to Step 10 (health display).
• y → proceed to Step 5b.

Step 5b — Download spec and plan:

python3 ~/.duplocloud/bin/duplo_ticket.py --get-spec

• If output is non-empty: save it to .duplocloud/spec.md.
• If output is empty: leave any existing .duplocloud/spec.md unchanged (no platform content yet).

python3 ~/.duplocloud/bin/duplo_ticket.py --get-plan

• If output is non-empty: save it to .duplocloud/plan.md.
• If output is empty: leave any existing .duplocloud/plan.md unchanged.

Tell the user which files were updated, then continue to Step 10.

Step 6 — List projects (filtered by workspace in state):

python3 ~/.duplocloud/bin/duplo_activate.py --list

Show the numbered project list to the user. Write out each numbered item clearly in your response text so the user can read the full list without expanding tool output.

Step 7 — Ask user to select project:

"Which project would you like to activate? (enter the number)"

python3 ~/.duplocloud/bin/duplo_activate.py --select <N>

(Replace <N> with the number the user provided.)

Step 8 — Show the confirmation output to the user.

Step 9 — Sync local files (new project — always overwrite):

python3 ~/.duplocloud/bin/duplo_ticket.py --get-spec

• If output is non-empty: save it to .duplocloud/spec.md.

python3 ~/.duplocloud/bin/duplo_ticket.py --get-plan

• If output is non-empty: save it to .duplocloud/plan.md.

Tell the user which files were written (or "no platform content yet" if both were empty). Continue to Step 10.

Step 10 — Show project health:

python3 ~/.duplocloud/bin/duplo_ticket.py --check-project

• If exit code is 3: tell the user "Project state missing — please re-run /duplo:activate_project." and stop.
• If exit code is 1: show the error and stop.
• If exit code is 0: parse the JSON and display a health summary:

Project: <project_name>

Artifact	Status
Spec	Approved / Draft (local file exists) / Not started
Plan	Approved / Draft (local file exists) / Not started

Determine artifact status independently for each artifact:

Spec status:

• action = "plan" or "none" → Approved
• action = "spec" and .duplocloud/spec.md exists → Draft
• action = "spec" and .duplocloud/spec.md does not exist → Not started

Plan status:

• action = "none" → Approved
• action != "none" and .duplocloud/plan.md exists → Draft
• action != "none" and .duplocloud/plan.md does not exist → Not started

Then ask the user whether to proceed:

• "spec" → ask: "Would you like to activate the spec ticket now? (y/n)"
  • y → run the activate_ticket skill inline (follow its steps from the beginning).
  • n → stop.
• "plan" → ask: "Spec is approved. Would you like to activate the plan ticket now? (y/n)"
  • y → run the activate_ticket skill inline (follow its steps from the beginning).
  • n → stop.
• "none" + has_execution_tasks = true → tell the user "Both spec and plan are approved. Execution tasks are ready." then ask: "Would you like to pick up an execution task now? (y/n)"
  • y → run the activate_ticket skill inline (follow its steps from the beginning, starting at Step 2b since has_execution_tasks is already known to be true).
  • n → stop.
• "none" + has_execution_tasks = false → tell the user "Both spec and plan are approved — waiting for the platform to generate execution tasks." and stop.