CLI DX: 'kelos logs' on a terminal-phase task gives no actionable guidance — doesn't suggest 'kelos get tasks <name> -d'

[kelos-bot]

🤖 Kelos User Agent @gjkim42

Problem

When a task has already completed (Succeeded or Failed) and its pod is gone, running `kelos logs ` returns:

```
Error: task "task-r8x2q" has no live pod (task phase: Failed)
```

or

```
Error: task "task-r8x2q" has no live pod (task phase: Succeeded)
```

This error is unhelpful because:

1. For failed tasks: The task's `status.message` often contains the actual failure reason (e.g., "pod exceeded active deadline"), but the user is left with no way to discover this from the error.
2. For succeeded tasks: The user may be trying to find their results (branch name, PR URL, cost) but gets a confusing dead-end error.
3. No next step is suggested: New users have no idea they should run `kelos get tasks -d` to see the failure message and outputs.

Root Cause

In internal/cli/logs.go (around line 58):

```go
if isTerminalTaskPhase(task.Status.Phase) {
return fmt.Errorf("task %q has no live pod (task phase: %s)", args[0], task.Status.Phase)
}
```

The error doesn't include the task's `status.message` or direct the user to the detail view.

Expected Behavior

The error message should be actionable. For example:

• For a Failed task: include `status.message` if set, and suggest `kelos get tasks -d` to see full details.
• For a Succeeded task: suggest `kelos get tasks -d` to view results (branch, PR URL, token usage).

Example improvement:

```
Error: task "task-r8x2q" failed (pod exceeded active deadline). Run 'kelos get tasks task-r8x2q -d' for full details.
```

or

```
Error: task "task-r8x2q" succeeded but has no live pod — logs are no longer available. Run 'kelos get tasks task-r8x2q -d' to view results.
```

Impact

This is a first-contact failure mode: a new user runs a task, it fails, they run `kelos logs` to debug, and they hit a wall with no guidance. This reduces confidence in the tool and increases friction during onboarding.

Scope

Small, self-contained change in internal/cli/logs.go — just improve the error message for the terminal-phase branch.

[kelos-bot]

🤖 Kelos Task Status

Task kelos-triage-791 has succeeded. ✅

[kelos-bot]

🤖 Kelos Agent @gjkim42

Kelos Triage Report

Kind: kind/feature
Priority: priority/important-longterm — Quality-of-life improvement for CLI error messages; not a bug since the command correctly reports the state, but the message lacks actionable guidance
Actor: actor/kelos — Self-contained change to a single error format string in internal/cli/logs.go with clear acceptance criteria

Status: STILL VALID

Evidence:

• Confirmed the issue at internal/cli/logs.go:58-59: the terminal-phase error returns only task %q has no live pod (task phase: %s) with no status.message content and no suggestion to run kelos get tasks <name> -d
• No merged PRs reference issue CLI DX: 'kelos logs' on a terminal-phase task gives no actionable guidance — doesn't suggest 'kelos get tasks <name> -d' #791
• No recent commits on main address this error message
• The code is unchanged and the reported behavior still exists

Duplicates found:

• CLI DX: 'kelos logs' gives unhelpful error for tasks in Waiting phase #716 (related but distinct — covers the Waiting phase error message, not the terminal/Succeeded/Failed phase addressed here)
• CLI UX: 'axon run --watch' doesn't show task results (PR URL, branch, cost) on completion #402 (tangentially related — covers --watch not showing results on completion, different command and flow)

Recommendation: KEEP OPEN