Use when a saved plan should be tested for cold-start executability before coding, especially for large tasks, weak-confidence plans, or plans whose early phases may be too vague for a fresh executor to follow without hidden context.
Scanned 5/27/2026
Install via CLI
openskills install Smithbox-ai/ControlFlow---
name: controlflow-executability-verifier
description: "Use when a saved plan should be tested for cold-start executability before coding, especially for large tasks, weak-confidence plans, or plans whose early phases may be too vague for a fresh executor to follow without hidden context."
---
# ControlFlow Executability Verifier
## Overview
Simulate the plan as if a fresh session had only the saved artifact and the repository.
This skill identifies whether the first tasks are specific enough to execute without
invisible assumptions.
Invoke this skill via `/controlflow-claude-code:controlflow-executability-verifier`.
## Workflow
1. Read the saved plan artifact first.
2. Simulate only the first few phases or tasks that matter for cold-start execution.
3. For each task, apply the 8-point checklist from
[references/executability-checklist.md](references/executability-checklist.md):
- what_clear
- where_clear
- how_clear
- inputs_defined
- outputs_defined
- dependencies_met
- verify_command_complete
- test_specifics_concrete
4. Walk through the task using the TDD walk-through from the checklist:
- open_file
- read_existing_code
- write_test_red
- run_test
- write_implementation_green
- run_test_again
- refactor
5. Stop at the first real blocker for a task and record exactly why it blocks execution.
6. Save the verdict to `plans/artifacts/<task-slug>/executability-verifier.md` using
[../../templates/executability-verifier-report-template.md](../../templates/executability-verifier-report-template.md).
7. Return a structured verdict:
- `PASS`
- `FAIL`
- `WARN`
- `ABSTAIN`
## Output Shape
- **Status**
- **Tasks Simulated**
- **Per-Task Checklist**
- **Walkthrough Summary**
- **Blocked Steps**
- **Recommendation**
- **Failure Classification** when needed
## Common Mistakes
- Simulating the plan with extra unstated context from memory.
- Continuing a task after a genuine blocker instead of stopping and recording it.
- Treating vague instructions as executable just because an experienced human could guess
the intent.
## References
- `references/executability-checklist.md`
- `../../templates/executability-verifier-report-template.md`
No comments yet. Be the first to comment!