Use to build or audit RTM links across requirements, design, validation, tasks, PRs, commits, releases, and evidence gaps.
Scanned 9/6/2026
Install to Claude Code
npx -y skills add BlueSkyXN/Codex-is-all-you-need --skill sdlc-requirements-traceability --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Sdlc Requirements Traceability?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/blueskyxn-sdlc-requirements-traceability)More formats (shields.io, HTML) on the badges page.
---
name: sdlc-requirements-traceability
description: Use to build or audit RTM links across requirements, design, validation, tasks, PRs, commits, releases, and evidence gaps.
metadata:
version: "0.5"
updated: "2026-07-26"
---
# Requirements Traceability
Use this workflow to create, update, or inspect a requirements traceability matrix.
Traceability connects intent to execution:
```text
Business / User / Product / Software Requirement
-> HLD / LLD / ADR / Domain Boundary / SPEC / Validation Plan / Handoff / Task
-> Test / Review / PR / Release
```
Traceability is strongest when SDLC materials exist. When no SDLC materials exist, traceability can still start from an issue, bug report, task ID, test failure, or direct-dev handoff. Do not block small development work solely because a full RTM is absent; create a traceability seed and mark the evidence level.
For lightweight SDLC-ADS work, use the operating-model contract as the default:
```text
REQ -> TASK -> VAL
ARCH / DOM / DEC -> TASK
ITEM -> TASK / VAL / Q
```
Do not require BRD/URS/PRD/SRS prefixes before a clear direct-dev, handoff-lite, or refactor task can proceed.
## Use when
- Requirements, specs, tasks, tests, PRs, or release notes must be connected.
- A feature has BRD, URS, PRD, SRS, NFR, HLD, LLD, ADR, Domain Boundary Map, SPEC, Validation Plan, or dev handoff artifacts.
- Work spans multiple tasks, modules, agents, or validation paths.
- You need to detect orphan requirements, orphan tasks, missing tests, or untracked changes.
- A change request modifies a prior requirement baseline.
- Dev needs requirement IDs to reference in task cards, commits, PRs, or review notes.
## Do not use when
- The task is a small direct code change with no need for formal tracking beyond issue and validation notes.
- The user only asks to implement, debug, test, or review code.
- No stable source material exists and the immediate task is safe to route to dev; create a task handoff instead.
- The request is to approve scope or release readiness. Recommend `$codex-next:sdlc-readiness-review` or use `dev-release-check` as appropriate.
## Inputs
Use what exists. Do not require all artifact families.
### Source artifacts
- BRD or business brief
- URS or user scenarios
- PRD or product requirements
- NFR matrix
- SRS
- HLD
- LLD
- ADR
- Domain Boundary Map
- SPEC slices
- validation plan, acceptance test matrix, smoke checklist, regression scope, or evidence expectations
- solution spec
- change control record
### Execution artifacts
- dev handoff
- task cards
- issue tracker items
- test plan or test cases
- validation command logs
- PRs, commits, branches, or release notes
- bug reports or incidents
### Direct-dev traceability sources
When SDLC artifacts are absent, trace from:
- issue ID
- bug ID
- user request
- reproduction step
- failing test
- file path or symbol
- task handoff ID
## Workflow
### 1. Choose traceability depth
Use the depth that matches the lane and risk.
| Depth | Use when | Matrix shape |
|---|---|---|
| `seed` | direct-dev, issue-backed, bugfix, early draft | Source -> TASK -> VAL |
| `standard` | contained feature delivery | REQ -> TASK -> VAL -> Test |
| `architecture` | architecture/domain-sensitive delivery | REQ + ARCH/DOM/DEC -> TASK -> VAL -> Test |
| `extended` | system, rebuild, multi-wave, or release-bound delivery | REQ + ARCH/DOM/DEC -> TASK -> VAL -> Test -> PR -> Release |
| `audit` | baseline, compliance, high-risk, or release-bound evidence | Same chain plus owner, status, evidence, change record |
Do not force `extended` or `audit` depth on every change.
### 2. Inventory source IDs
Collect stable IDs from available artifacts.
Use these prefixes:
| Prefix | Meaning |
|---|---|
| `REQ` | Requirement from requirements package or scope decision table |
| `DEC` | Durable decision |
| `ARCH` | Architecture constraint |
| `DOM` | Domain, ownership, data, or dependency constraint |
| `TASK` | Dev task |
| `VAL` | Validation item |
| `Q` | Open question or blocker |
| `ITEM` | Midstream intake item |
If an input has no ID, create a lightweight local ID and mark it as generated. Use only the prefixes above.
### 3. Normalize requirement rows
Each row should represent one traceable item or one traceable chain.
Base columns:
```markdown
| Trace ID | Source | Requirement | Type | Priority | Downstream | Task | Validation | Status |
|---|---|---|---|---|---|---|---|---|
```
Extended columns:
```markdown
| Trace ID | Source | REQ | ARCH/DOM/DEC | TASK | VAL | Test | PR/Commit | Release | Owner | Status | Evidence | Change |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
```
Use compact rows. A traceability matrix should help dev and review, not become an unreadable archive.
### 4. Link source to software-facing requirements
Map source materials to `REQ`, `TASK`, and `VAL`.
Examples:
```text
REQ-001 -> TASK-001 -> VAL-001
ARCH-001 -> TASK-002
DOM-001 -> TASK-003
Issue-123 -> TASK-001
Bug-456 -> TEST-REG-001
```
Rules:
- One upstream item may map to several downstream requirements.
- One downstream task may satisfy several requirements only if the task scope is still coherent.
- Mark inferred links explicitly.
- Do not hide many-to-many relationships by collapsing unrelated items.
### 5. Link requirements to architecture/domain artifacts and SPEC slices
For each `REQ` or task item, identify needed architecture/domain and specification evidence:
| Requirement | Needed architecture/domain constraint | Needed specification detail | Status |
|---|---|---|---|
| `REQ-001` | `ARCH-001`, `DOM-001`, `DEC-001` | `20-规格.md` API section | present / missing / not applicable |
If no specification detail is required, state why.
If no HLD, LLD, ADR, or Domain Boundary Map is required, state why when the work is architecture-sensitive.
### 6. Link validation plan to tasks and tests
Each implementation task should have a validation path. When a Validation Plan or acceptance matrix exists, link tasks to the relevant validation IDs before linking to executed test evidence.
| Task | Validation Plan item | Test or check | Type | Status |
|---|---|---|---|---|
| `TASK-001` | `VAL-001` | `pytest path/test_name` | unit / integration / e2e / manual / smoke / NFR | planned / passed / failed / not run |
Do not mark validation as passed unless evidence exists.
### 7. Record execution evidence when available
When dev has started or finished, add:
- branch name
- PR number or link
- commit SHA if relevant
- test command and result
- review verdict
- release or deployment reference
If execution evidence is not yet available, leave it blank or mark `pending`.
### 8. Detect traceability gaps
Report gaps by type:
| Gap | Meaning |
|---|---|
| orphan source | requirement has no downstream SRS/task |
| orphan task | task has no source or issue link |
| missing validation | task or requirement has no test/validation |
| missing validation plan | feature, system, release, NFR, architecture, or domain-sensitive work lacks explicit proof-of-correctness plan |
| missing SPEC | requirement needs UI/API/Data/etc. spec but none exists |
| missing architecture artifact | requirement needs HLD, LLD, ADR, Domain Boundary Map, or Directory SPEC but none exists |
| boundary conflict | task or SPEC appears to violate ownership, dependency, or forbidden access rules |
| stale link | target artifact is superseded |
| unapproved change | task implements changed scope without change record |
| direct-dev risk | implementation proceeds from issue/task without formal SDLC materials |
Not every gap blocks dev. Classify severity:
```text
blocker
important
watch
accepted
```
### 9. Support direct-dev traceability
For a direct-dev task, produce a seed matrix:
```markdown
| Trace ID | Source | Task | Validation | Risk | Status |
|---|---|---|---|---|---|
| TRACE-001 | Issue-123 | TASK-001 | pytest path/test_name | low | planned |
```
This preserves evidence without pretending a full SDLC chain exists.
## Output
Return or write one of these outputs.
### Traceability seed
```markdown
# Traceability Seed: <Change Name>
| Trace ID | Source | Task | Validation | Risk | Status |
|---|---|---|---|---|---|
```
### Standard RTM
```markdown
# Requirements Traceability Matrix: <Feature / System>
## 1. Scope
## 2. Artifact Sources
## 3. Traceability Matrix
## 4. Gap Report
## 5. Execution Evidence
## 6. Change Notes
```
### Gap report
```markdown
# Traceability Gap Report
## Blocking gaps
## Important gaps
## Accepted gaps
## Recommended next action
```
## Validation
Before declaring traceability ready, check:
- Every P0/P1 SRS requirement has a downstream task or explicit deferment.
- Every implementation task has a source or direct-dev source note.
- Every requirement or task has a validation plan or explicit accepted gap.
- NFRs have validation methods, not only statements.
- Validation Plan items link to tasks and execution evidence where a validation plan exists.
- Architecture-sensitive requirements point to HLD, LLD, ADR, Domain Boundary Map, or an explicit accepted gap.
- SPEC-dependent requirements point to the relevant SPEC slice or missing-SPEC gap.
- Change-related rows point to existing change-control records when baseline is affected.
- Status values are consistent.
- Direct-dev work is not misrepresented as full SDLC-backed work.
## Boundaries
- Do not approve requirements.
- Do not execute tests or claim validation success without evidence.
- Do not block safe direct-dev work solely because a full matrix is absent.
- Do not create fake BRD, URS, PRD, SRS, or SPEC IDs to make the matrix look complete.
- Do not use traceability as a substitute for `sdlc-dev-handoff-planning`.
- Do not overwrite existing tracking conventions without user approval.
## Handoff
Recommend the smallest next step. For an explicit-control workflow, return its
exact `$codex-next:<skill-name>` command and stop; do not begin it here.
Route findings:
| Finding | Next step |
|---|---|
| source not ready | `$codex-next:sdlc-requirements-workflow`, `sdlc-prd-workflow`, or `sdlc-srs-workflow` |
| HLD missing | `sdlc-hld-workflow` |
| LLD missing | `sdlc-lld-workflow` |
| ADR missing | `sdlc-architecture-decision-record` |
| domain boundary missing | `sdlc-domain-boundary-modeling` |
| SPEC missing | `sdlc-spec-slice-writer` |
| validation plan missing | `sdlc-validation-plan-workflow` |
| handoff missing | `sdlc-dev-handoff-planning` |
| baseline changed | `$codex-next:sdlc-change-control` |
| readiness uncertain | `$codex-next:sdlc-readiness-review` |
| implementation ready | dev skill such as `dev-spec-driven-implementation`, `dev-bugfix`, or `dev-test-strategy` |
| direct-dev safe but untracked | create traceability seed and proceed to dev |
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!