danmestas SPDD Alignment
Lock intent before design or code. This skill turns raw input into a shared, testable problem statement.
Announce at start: "I'm using the spdd-alignment skill to lock intent before writing the prompt or code."
Hard gate
Do not invoke implementation skills, edit product code, or generate a REASONS Canvas until the alignment questions below have concrete answers. If a question is answerable from repository context, answer it by inspection rather than asking the user.
Alignment checks
Business value
- What user/business problem is being solved?
- What outcome would prove the change mattered?
- Which behavior is explicitly out of scope?
Domain language
- Define the key nouns and verbs in the user's language.
- Identify same-word/different-meaning risks.
- Prefer existing repository terminology unless the requirement intentionally changes it.
Acceptance criteria
- Convert vague success statements into Given/When/Then or equivalent observable checks.
- Cover happy path, important boundaries, and invalid inputs.
- Separate business acceptance from implementation detail.
Constraints and dependencies
- Identify upstream systems, unfinished modules, migration order, data compatibility, security, performance, compliance, and operational constraints.
- Preserve legacy behavior unless the aligned requirement explicitly changes it.
Output
Produce an alignment.md section or artifact with:
- Problem statement.
- Business value.
- Scope in/out.
- Domain glossary.
- Acceptance criteria.
- Constraints and dependencies.
- Open questions that are truly blocking.
If blockers remain, do not proceed to spdd-analysis until they are resolved or explicitly deferred by the user.