sdlc-design
Turn an accepted .sdlc/feature-slug/intent.md into a policy-constrained, reviewer-approved spec.md when asked to design or specify an approved feature. Do NOT use for raw idea discovery, coding, test execution, PR review, deployment, or production monitoring.
- License 许可
- Apache-2.0
- 最近更新
- 2026-11-14
- Stars
- ★ 1.6k
- Forks
- 491
01 前言 Front Matter
- name
- sdlc-design
- description
- Turn an accepted .sdlc/feature-slug/intent.md into a policy-constrained, reviewer-approved spec.md when asked to design or specify an approved feature. Do NOT use for raw idea discovery, coding, test execution, PR review, deployment, or production monitoring.
03 SKILL.md 正文
SDLC Design
Run Stage 2 of 6 in the AI-Native SDLC playbook described in docs/PLAYBOOK.md.
Purpose
Generate a testable specification from accepted intent while making policy collisions and unresolved decisions visible.
The primary value of this stage is the Flagged concerns section. Never solve or hide a concern silently.
Entry gate
Locate .sdlc/<slug>/intent.md.
Proceed only if it exists and declares Status: accepted.
If the artifact is missing, still draft, or lacks explicit acceptance, route the
work back to sdlc-plan and stop.
Read the entire accepted intent before generating requirements or a design.
Discover governing policy
Discover available organization skills and relevant project instructions before designing. Look specifically for brand, security, compliance, and data-handling skills or policies.
List each discovered skill or policy that constrains the design.
If none are found, log that no organization skills were available. Do not pretend that an undiscovered policy has been applied.
Read the applicable guidance and constrain the proposed design by it.
Procedure
- Read the accepted
intent.mdin full. - Discover and read applicable organization skills and project policies.
- Log the skills or policies found, or log that none were found.
- Convert desired outcomes into numbered requirements that can be verified.
- State clear non-goals to protect the accepted scope.
- Design the architecture, flow, data model, and interfaces within policy.
- Describe API, schema, persistence, and contract changes when applicable.
- Identify every collision among intent, policy, constraints, and open questions.
- Mark every concern as
blockingornon-blockingand explain why. - Describe how the final implementation will be verified.
- Write
.sdlc/<slug>/spec.mdand show the complete draft to the user.
Use plain language. Do not replace a requirement with an implementation preference.
Required artifact
Start with this contract structure. Add Status: draft directly below the title
while the document awaits review, then change it to Status: accepted only after
explicit approval.
# Spec: [feature name] (from intent.md [date])
## Source intent
[one-line link to or quote from the accepted intent.md]
## Requirements
1. [testable requirement]
2. [testable requirement]
[etc.]
## Non-goals
[explicitly what this change does NOT do]
## Design
[architecture, flow, data model, interfaces]
## Data and interfaces
[API/schema changes, persistence, contracts]
## Flagged concerns
[CRITICAL: each blocking/non-blocking issue, never silently resolved]
- [concern]: [why it matters], [blocking|non-blocking]
## Verification strategy
[how we'll prove the spec worked]
Keep the contract headings and the concern format unchanged.
Flagged concerns rules
A concern is blocking when the specification cannot safely proceed without a decision or approval. Example: a new database table requires operations approval.
A concern is non-blocking when it is valuable to resolve but the scoped change can ship without it. Example: the first release does not cover every locale.
Flag both kinds. Do not downgrade a blocker because a workaround seems likely.
Do not silently resolve an open question, policy collision, exception, or risk.
When a concern needs a decision, state the decision needed and its owner if known.
Exit gate
Show the entire spec, including every flagged concern, then stop and ask exactly:
Reviewer, do you accept this scope and the flagged concerns?
Accept only explicit human approval. Do not code, reinterpret approval, or proceed from an implied response.
After explicit approval, change the status to accepted in spec.md and commit
with Status: accepted in the commit message.
If the reviewer requests changes, revise the draft, preserve or update concerns, and repeat the exit gate.
Handoff
Hand off the accepted .sdlc/<slug>/spec.md to sdlc-build.
Include the policy sources consulted and every unresolved non-blocking concern.
Do not decide a blocker on behalf of the reviewer or organization.
04 同仓库更多 Skills
worktable-review-with-annotations
Create, inspect, reply to, update, or resolve situated Worktable annotations. Use when feedback belongs to a Doc section, block, text range, or HTML Doc, or when the user asks to address existing review feedback and coordinate a requested artifact edit.
worktable-find-and-synthesize-context
Find, assess, and synthesize context from Worktable Docs and Records. Use when the user asks what a workspace contains, wants workspace-backed research, needs current facts reconciled across artifacts, or requests a cited synthesis without changing workspace state.
worktable-create-or-update-docs
Create, edit, rename, or maintain narrative Worktable Docs. Use for plans, decisions, briefs, notes, research syntheses, and other durable prose when the user requests a new artifact or a safe update to an existing one.
worktable-create-or-manage-records
Organize repeated structured information as durable Worktable Records. Use for research datasets, inventories, directories, pipelines, trackers, and other collections whose items need identity, fields, filtering, links, validation, or individual updates.
worktable-create-or-manage-interactive-html
Create or improve durable visual and interactive artifacts in Worktable. Use for dashboards, trackers, calculators, mini apps, slide presentations, and other requests that benefit from a polished HTML Doc.
worktable-collaborate-threads
Read, answer, and hand off messages through durable Worktable threads with users or other authenticated participants. Use when the user asks to check addressed messages, reply to a participant, continue a thread, request another agent's help, wait briefly for a response, or watch for comments and messages sent to the agent.