Repository navigation
[Feature]: Add /speckit.guide for dry-run implementation planning #3057
Description
Problem Statement
While /speckit.clarify helps reduce ambiguity in the feature specification before planning, there is currently no equivalent step for validating implementation scope after task generation and before execution.
This leaves a gap where /speckit.implement may exceed the intended scope by acting on future tasks or making architectural assumptions that were not explicitly reviewed beforehand.
This creates risks such as:
Scope creep across planned phases
Unintended architecture changes
Hallucinated dependencies or missing interfaces
Higher review and rollback costs
Proposed Solution
Introduce a new optional command: /speckit.guide.
This command would run after /speckit.tasks and before /speckit.implement, generating a guide.md file that acts as a dry-run implementation plan.
The generated guide would describe:
- Which files are expected to be modified
- Which interfaces or services may be affected
- Required dependencies or prerequisites
- Execution order of tasks
- Scope boundaries for the current phase
- Potential risks or stop conditions
This would provide developers with a reviewable implementation artifact before execution, reducing the chance of scope drift and improving architecture control.
Proposed workflow:
/speckit.specify
→ /speckit.clarify
→ /speckit.plan
→ /speckit.tasks
→ /speckit.guide
→ Human review
→ /speckit.implement
This would not replace /speckit.clarify, but complement it by extending the same validation principle into the implementation phase.
Alternatives Considered
Yes, I considered the following alternatives:
-
Using
/speckit.clarify- This helps resolve ambiguity in the feature specification before planning.
- However, it does not address implementation scope validation after task generation.
-
Improving
/speckit.implementdirectly- Scope validation logic could be embedded into
/speckit.implement. - However, this would mix planning/review concerns with execution, reducing workflow transparency.
- Scope validation logic could be embedded into
-
Manual
guide.mdcreation- This is the current workaround I use.
- It provides better control, but it is manual, inconsistent, and not integrated into the Spec Kit workflow.
I believe a dedicated /speckit.guide command is the cleanest solution because it introduces a reusable, explicit review layer between task generation and implementation.
Component
Agent integrations (command files, workflows)
AI Agent (if applicable)
None
Use Cases
No response
Acceptance Criteria
No response
Additional Context
No response
Activity
himanshu-commits commentedon Aug 8, 2026
I'd like to pick this up, but wanted to check direction first.
Four of the six items listed for
guide.mdare largely covered by/speckit.taskstoday — the tasks template already requires exact file paths per task and generatesDependencies & Execution Order,Parallel Opportunities, andImplementation Strategy. The two that genuinely have no home are scope boundaries for the current phase and stop conditions. Those are real gaps:/speckit.analyzesits in the same slot but reports to the console rather than persisting a reviewable artifact, and/speckit.implementhas no scope fence at all.So my questions:
/speckit.analyzeget the same benefit more cheaply? A new command means registration across the core command lists, per-agent generated files, the command-set assertions in the integration tests, hook docs, and the README/docs tables.guide.mddeliberately not restate what's already intasks.md, and instead be a short scope contract that references it?Happy to implement either. I can draft the command template and run it against a real feature's
tasks.mdfirst so we can compare the two files before committing to the plumbing.mnriem commentedon Aug 10, 2026
Feel free to take it and publish it as a community extension
github-actions commentedon Oct 10, 2026
This issue has been automatically marked as stale because it has had no activity for 60 days. It will be closed in 30 days unless there is a substantive update. If the problem still applies, please provide updated context such as whether it is still reproducible, the affected version or environment, its current impact, and any new reproduction details. A comment that only states that the issue is "still relevant" does not give maintainers enough information to reconsider it.