When a slash-skill is registered but its payload fails to load, the CLI treats the failure string as a successful command payload and queues it into _pending_input instead of surfacing a user-facing error.

• agent/skill_commands.py:311-313 — build_skill_invocation_message() returns a truthy error string like [Failed to load skill: plan] A runtime skill-load failure is converted into fake conversational input. The model receives an error sentinel as if it were a real user/system payload, while the CLI prints a success message (Plan mode queued...). This is a silent failure mode and can confuse both the user and the model. If a skill payload fails to load, the CLI should show an error and avoid queueing any payload. Make build_skill_invocation_message() return a structured failure result (or None plus logged error) instead of a truthy string, and update /plan plus generic skill dispatch to reject error sentinels explicitly.

TL;DR

Modify build_skill_invocation_message() to return a structured failure result or None with a logged error, and update the /plan and generic skill dispatch paths to reject error sentinels explicitly.

Guidance

• Update build_skill_invocation_message() in agent/skill_commands.py to return a structured failure result, such as a dictionary with an error message, or None with a logged error.
• Modify the /plan path in cli.py to check for a specific error sentinel or a structured failure result instead of just checking truthiness.
• Update the generic skill command path in cli.py to reject error sentinels explicitly, ensuring that only valid payloads are queued.
• Verify the changes by running the minimal reproduction script and checking that the CLI correctly handles skill payload failures.

Example

# agent/skill_commands.py
def build_skill_invocation_message():
    # ...
    if failure:
        # Log the error and return None or a structured failure result
        logging.error("Failed to load skill: plan")
        return None
    # ...

# cli.py
def _handle_plan_command(self, command):
    # ...
    payload = build_skill_invocation_message()
    if payload is None:
        # Handle the error case explicitly
        print("Error: Failed to load skill")
    else:
        # Queue the valid payload
        self._pending_input.put(payload)

Notes

The suggested changes assume that the build_skill_invocation_message() function can be modified to return a structured failure result or None with a logged error. Additional error handling may be necessary to ensure that the CLI correctly handles all possible failure cases.

Recommendation

Apply the suggested workaround by modifying build_skill_invocation_message() and updating the /plan and generic skill dispatch paths to reject error sentinels explicitly, as this will allow the CLI to correctly handle skill payload failures