Start with Reqode
REQODE FEATURES

Change Management

Keep every product change scoped, traceable, and ready to implement.

A Change Request connects why the product should change with its project branch, affected specification revisions, implementation scope, related AI work, and Issue. So people and coding agents work from the same context.

Start with Reqode
Start with Reqode

A product change is more than a ticket

A ticket can describe a task. It rarely preserves the exact product context needed to analyze, implement, verify, and revisit the change months later.

DISCONNECTED CHANGE

The reason, specifications, and implementation live apart

  • The request is described in a ticket or prompt.
  • Affected specifications are reconstructed during implementation.
  • AI research and decisions disappear into separate sessions.
  • Reviewers must rediscover which branch, code, and product behavior apply.
REQODE CHANGE REQUEST

One stable context follows the change

  • The intent and working project branch are explicit.
  • Exact affected specification revisions define the scope.
  • Related AI work, Issues, subsystems, and software units stay connected.
  • One CR-* key gives people and agents a shared starting point.

One change. One implementation context.

A Change Request packages the information needed to understand what is changing, where the work happens, and which product and code areas require attention.

Purpose and context

Give the change a stable CR-* key, a clear name, and a Markdown description that preserves the motivation and expected outcome.

Project branch

Define the project branch where the change is planned and implemented, including isolated feature, migration, and hotfix work.

Affected spec revisions

Connect the exact Requirements, User Interfaces, Data Entities, and API Operations affected by the change—including their revision and branch context.

Implementation impact

Derive affected subsystems from the selected specs and inspect the related software units and alignment signals in a read-only Code view.

AI and Issue context

Keep linked AI Task summaries and the originating Issue close to the change, without treating a temporary AI answer as product truth.

Clear lifecycle

Work with an Active Change Request, close it into a read-only record when the change is complete, and reopen it when more work is required.

FROM INTENT TO CODE

See the full scope without reading every artifact

The General view explains the change and lists the specification revisions that define its product scope. The Code view derives the connected software units from those affected specs.

Reviewers can move from a change to a historical spec revision, the current branch-effective artifact, or a related unit and its alignment evidence.

The result is a navigable impact map, not a manually maintained checklist.

EXAMPLE CHANGE CONTEXT
Add enterprise SSO
CR-184ACTIVEFEATURE-AUTH
Affected Specifications
REQ-42 · Enterprise authentication
REV. 3Created in branch
UI-18 · Sign-in screen
REV. 2Forked in branch
API-27 · Start SSO session
REV. 4
Derived Code Impact
U-214 · AuthenticationService
REQ-42ALIGNMENT
U-228 · SignInController
UI-18API-27

A controlled path from decision to delivery

Use only as much structure as the change needs, while keeping the order of work and the source of truth explicit.

1
CAPTURE

Describe the intent

Record why the behavior must change, the expected outcome, and any implementation guidance that belongs to the change.

2
ISOLATE

Choose the project branch

Use Main for direct work or a project branch when the change needs an isolated effective view before it joins the main product model.

3
SCOPE

Connect affected specs

Add specification revisions manually, or review and Apply a linked AI Analyst result so its final post-Apply revisions become part of the scope.

4
ASSESS

Inspect implementation impact

Review derived subsystems, software units, source-spec relations, and alignment markers before deciding where code must change.

5
HAND OFF

Give the agent one key

Pass the CR-* key to a coding agent. Reqode MCP resolves the affected artifacts in the correct branch and exposes related implementation context.

6
VERIFY

Check the result and close

Verify specifications, units, code, and tests. Close the Change Request into a read-only record when the work is complete.

Useful throughout the product lifecycle

The same structure works for planned features, urgent corrections, architecture changes, and AI-assisted discovery.

NEW FEATURE

Add enterprise SSO

Starting pointA product decision needs coordinated changes across authentication requirements, UI, and API behavior.
Change contextA feature branch, the affected spec revisions, linked AI Analyst work, and the derived authentication units.
OutcomeProduct, engineering, QA, and the coding agent implement against the same accepted scope.
PRODUCTION CORRECTION

Fix incorrect tax calculation

Starting pointAn Issue identifies behavior that differs from the expected calculation rules.
Change contextA hotfix branch, the linked Issue, affected data and API specs, and the units responsible for calculation.
OutcomeThe team can decide whether the code, the specification, or both must change—then preserve that decision.
ARCHITECTURE EVOLUTION

Replace the authentication provider

Starting pointA technical migration affects several user journeys and implementation units.
Change contextThe relevant product behavior stays explicit while MCP adds current subsystem and architecture guidance for implementation.
OutcomeThe migration remains anchored to product intent instead of becoming a code-only refactor.
AI-ASSISTED CHANGE

Turn a policy change into reviewable specs

Starting pointThe business request is clear, but the affected requirements, data, API, and UI behavior are not.
Change contextAI Analyst prepares a structured delta, the team reviews it, and Apply links the final affected revisions to the Change Request.
OutcomeThe coding agent receives reviewed product context rather than an unbounded AI conversation.
CODING AGENT HANDOFF

Give the agent a change key, not a context dump

A coding agent can start from a Change Request key while Reqode MCP resolves the affected artifacts in its branch, guides the agent toward related software units and files, and keeps current specifications above temporary instructions or AI summaries.

If the change context conflicts with current product truth, the agent is instructed to report the conflict before implementation.

Use Reqode MCP as the source of truth.
Start from CR-184.
Resolve affected artifacts in its project branch.
Inspect related software units and repository files.
Report conflicts before implementation.
FREQUENTLY ASKED QUESTIONS

Where Change Requests fit

Change Requests do not replace specifications, Issues, or AI Tasks. They connect those concerns around a concrete product change and its implementation scope.

How is a Change Request different from an Issue?

An Issue captures an idea, concern, bug, or other reason for attention. A Change Request packages the product and implementation scope chosen to address a change. An Issue can be connected to one or more Change Requests when it leads to separate implementation paths.

Does every change need a Change Request?

No. A narrow, already understood change can start from an exact specification or software unit key. A Change Request is most useful when work crosses artifacts, needs branch isolation, involves AI analysis, or benefits from a durable handoff and audit context.

Can AI update affected specs automatically?

AI Analyst prepares a structured result for review. When the team applies a linked result in the same project branch, Reqode can add the final post-Apply specification revisions to the Change Request. Ordinary specification edits outside this flow are not silently added.

What happens when a Change Request is closed?

A closed Change Request moves to the Archive and becomes read-only while its affected specs and linked context remain available. A user with management permission can reopen it if more work is required.

Continue through the connected workflow

Explore the Reqode capabilities that prepare the change, deliver it to an agent, and verify the result.

AI Analyst Workflow

Turn product ideas into reviewable specification changes before implementation begins.

MCP Context

Give coding agents current product, branch, unit, file, and architecture context.

Living Documentation

Keep actual product intent and its historical revisions available in one model.

Continuous Verification

Detect drift, preserve evidence as Findings, and close the loop after implementation.

Start with one controlled change

Choose one product change, connect its affected context, hand one stable key to the coding agent, and verify the result against the same source of truth.

Start with Reqode
Start with Reqode