This skill is mandatory only on the task-file path of the
spec-loop-plan-task workflow. It is not used for chat-only planning
path work kept in chat.
Before work: read spec-loop-plan-task/SKILL.md, follow all files it
requires, and apply the full spec-loop-plan-task bundle as shared
convention guidance. Then read
execution-approval-guidance.md.
Update the same active task file for user-facing review readiness. Do not create a second task artifact or a new pre-execution task status.
Use execution-approval-guidance.md as the authoritative source for readiness checks, diagram-first preparation, duplication removal, user-facing review gating, and approval-seeking output.
Follow execution-approval-guidance.md for the allowed user-facing outcomes and response rules.
If readiness checks find directly fixable task defects, update the active task in place and re-run the readiness checks before producing user-facing output, except for structural decomposition changes handled specially by execution-approval-guidance.md.
If readiness checks find issues that can be resolved from existing evidence, update the active task in place and re-run the readiness checks before producing user-facing output.
If readiness checks show that task-file compaction is needed, use spec-loop-compact-task-file/SKILL.md and then resume this skill.
If any important open decision remains and it is user-preference-sensitive, needed to choose the next path, or could materially change scope, constraints, design, or test specification, do not end with a blocker list. Invoke spec-loop-clarify-task/SKILL.md in the same turn, then resume this skill after clarification before seeking execution approval.
After task-file execution approval:
- implementation work is governed by spec-loop-implementation-flow/SKILL.md;
- investigation work stays under the
spec-loop-plan-tasktask-file path: perform only the approved investigation, record final output inFindings, satisfy anyTest specification, and move the task or subtask toreview.
If no active task file exists, if the current increment is unclear, or
if chat-only planning is in use, do not use this skill. Continue task
work under the spec-loop-plan-task bundle instead.