Design SolutionSAFE

A collection of Agent Skills Standard and Best Practice for Programming Languages, Frameworks that help our AI Agent follow best practies on frameworks and programming laguages

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: design-solution
description: "Turn an approved PRD or implementation goal into SRS/FRS technical requirements (How), architecture, contracts, and verification decisions."
metadata:
  internal: true
  triggers:
    keywords:
    - design solution
    - workflow
---
# Design Solution Skill

> [!IMPORTANT]
> Turn an approved PRD or implementation goal into SRS/FRS technical requirements (How), architecture, contracts, and verification decisions.

Optional args: slug=<feature>, ticket=<id/url>, mode=interactive|autonomous|channel, channel=<id>, auto_continue=true|false, profile=business|hybrid|technical.

## Instructions

When the user asks to perform this workflow, execute the following steps:


# Design Solution Workflow (SRS/FRS / How)

Goal: Produce a build-ready technical design with explicit boundaries, contracts, risks, and tests.

## Steps

1. Load inputs:
   - Load baseline SRS/FRS section, `common-software-requirements`, `common-decision-discipline` (option cards, evidence ledger), PRD or ticket, implementation plan, matched framework skills, architecture docs, trace source `BRD-OBJ-* -> REQ-* -> AC-*`, and `common-product-requirements/references/requirements-modeling.md` from the owning package when modeling guidance applies; report excluded or unavailable resources without guessing.
2. Define architecture:
   - Name bounded contexts, module/data owners, migration needs; select technical contracts/views without inventing product scope.
   - Consume or create HLD trace: requirements, audience, scope, shaping constraints, lifecycle status, ownership, failure domains, decisions.
   - Define dependency direction, component RACI, and sync, async, or hybrid communication.
   - Define early mock/schema contracts for parallel start; create LLD contracts only for components/flows needing them. Select `container`, `sequence`, or other views by question through `common-architecture-diagramming` or `requirements-modeling.md` (`evidence` cites SRS/system-design doc, `metric` from stated NFR thresholds). Views clarify interaction/lifecycle/order; diagrams not mandatory when prose or tables are precise enough; never fabricate metrics.
3. Define contracts:
   - Functional flows (FRS): user/system steps, inputs/outputs, validations, and error states.
   - For complex flows, use one actor, one goal, one session; split normal course from alternatives and exceptions.
   - Requirement cards: statement, priority, status, source, behavior, NFRs, measurement, and verification lane.
   - API inputs/outputs and interface contracts (OpenAPI/Protobuf).
   - Events/jobs and async guarantees (at-least-once, idempotent).
   - Storage shape, ownership, retention, and migration rules.
   - Security, permission, and privacy checks.
   - Carry `REQ-* -> HLD-* -> CMP-* -> LLD-* -> VER-*` IDs into requirement cards and interface contracts. LLD is the low-level-design lane: specify ownership, consistency, ordering/idempotency, adverse timeline, recovery, and verification for each chosen component; do not force every component into a diagram.
   - NFR thresholds for performance, reliability, and scalability.
4. Plan verification:
   - Unit, integration, E2E, visual, mobile, security, and migration checks.
   - Failure mode analysis for dependencies, fallbacks, retries, and rollback/degradation.
   - Save technical requirements to `docs/srs/srs-[slug].md` when file writes are allowed.
   - Record evidence in `docs/srs/srs-walkthrough-[slug].md`.
5. Record ADR:
   - Write one concise ADR when architecture or public contract changes.
   - Continue when patterns are inferable; return BLOCKED for cross-team contracts, migrations, permissions, or NFR uncertainty.
   - Route next step to `implementation-readiness` or `dev-fix`.

## Runtime Contract

- Use after PRD or when implementation is approved but architecture and contracts are not explicit.
- Required inputs: PRD or ticket plus enough context to define contracts and verification.
- Return BLOCKED only for cross-team contracts, migrations, permissions, or NFR uncertainty.

## Handoff Payload

- `slug`, SRS path, requirement trace, architecture decisions, diagram paths, contracts, data/migration plan, NFR thresholds, verification matrix, ADR, outcome report, next workflow.

## Blocking Questions

- Ask max 3 at a time with a recommended default and 2-3 options.

## Output Template

```md
# Technical Design (SRS/FRS): [Name]
## Context
## Requirement Trace (BRD -> PRD -> SRS)
## Architecture & RACI
## Selected Views (optional; prose or tables allowed)
## Functional Flows (FRS)
## Parallel Readiness (Mocks/Schemes)
## Requirement Cards
## Contracts (API/Events)
## Data And Migration
## NFR Thresholds & Measurement
## Security And Privacy
## Failure Mode Analysis (FMA)
## Verification Plan & Evidence Matrix
## ADR

## Outcome Report
{schema_version: 1, run_id: "[run-id]", slug: "[slug]", workflow: design-solution, feature_status: design_ready, started_at: "[timestamp]", completed_at: "[timestamp]", requirement_trace: {brd_objectives: [], requirements: [], acceptance_criteria: [], srs: []}, completed_evidence: [], missing_evidence: [], decision_needed: [], recommended_next_workflow: implementation-readiness, cost: {source: unavailable}, agent: {identity: "[agent-identity]", model: "[model]"}}

## Next Workflow
implementation-readiness | dev-fix
## Cost Report
Call `get_session_cost(workflow="design-solution")` before final handoff.
```

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

What the source does

Filesystem

none-observed

Network

none-observed

Shell

none-observed

Dependencies

pinned

Secrets in source

none-found

Findings (0)

No findings outside the package's declared scope.

Gates applied: no_behavioural_pass.