Sigil: Invoke Example Runner

The short legacy commands are discovery bridges into this sigil:

• invoke-example-next resolves next,
• invoke-example-run <selection> resolves an exact task ID, template/complexity pair, or next.

• select exactly one prompt,
• preserve the prompt's task ID, template target, complexity, and user request,
• route through canonical invoke contracts and the active native skill surface by default,
• save a user-facing output file,
• verify Dispatch Spec technique trace evidence,
• verify Distill validation evidence for plan, full, and validate,
• preserve task-session as the next owner for bounded execution,
• avoid marking deferred modes as implemented,
• report any block or flag result honestly,
• leave canonical invoke contracts unchanged while running examples.

• claiming a prompt was run when only selected,
• overwriting an existing output unless the user requested a rerun,
• bypassing the active native invoke skill package,
• treating legacy command compatibility as native authoring readiness,
• editing canonical invoke contracts while running examples,
• accepting an output that lacks Dispatch trace or required Distill validation evidence,
• marking task execution complete when Invoke only emitted a Task Session handoff,
• replacing missing outputs with generic placeholder text.

## Invoke Example Runner Result

- Selection: <task-id>
- Prompt: <path>
- Output: <path>
- Execution surface: native invoke | legacy command | legacy CLI
- Invoke mode: define | design | plan | handoff | refresh | full | validate
- Invoke status: pass | flag | block
- Dispatch trace: pass | flag | block
- Distill validation: pass | flag | block | skipped | not_applicable
- Task-session handoff: ready | blocked | not_applicable
- Saved output: yes | no
- Next unrun: <task-id | none>
- Follow-up: <items | none>