---
name: repl-tasks
description: "Generate TASKS.md decomposing implementation into ordered, verifiable tasks. Use when the user asks to create TASKS.md or decompose work. (TASKS.md 작성을 요청받으면 사용)"
---

Generate TASKS.md.

Follow AGENTS.md.

Assume PRODUCT_SPEC.md, TECH_STACK.md, and ARCHITECTURE.md
already exist and are correct.

Stop and report instead of generating TASKS.md only if:

- the documents conflict and the conflict cannot be resolved by the priority rules in AGENTS.md
- PRODUCT_SPEC.md contains unresolved product-scope ambiguity or contradiction
- a required behavior is undefined in PRODUCT_SPEC.md and cannot be derived
  without inventing product behavior

Do NOT stop merely because any document contains [UNVERIFIED].

The purpose of TASKS.md is to define:

What should be implemented next, in dependency order.

TASKS.md is an execution checklist for AI implementation,
not project management.

Do not include estimates, owners, dates, or priorities.
Do not create any other document.

Do not describe architecture.
Do not describe technologies.
Do not describe tech stacks.
Do not describe implementation details.
Do not reference specific classes.
Do not reference specific files.
Do not reference specific libraries.

You may reference documents, document sections,
and component names defined in ARCHITECTURE.md.
Do not restate their content.

## Scope rules

PRODUCT_SPEC.md defines the complete current build scope.

Do not require PRODUCT_SPEC.md to distinguish MVP from Future.
Do not invent an MVP/Future classification.

Every buildable requirement stated in PRODUCT_SPEC.md is in the current scope.
Features excluded by PRODUCT_SPEC.md MUST NOT appear as tasks.

Do not create tasks for anything PRODUCT_SPEC.md does not state.

## Task rules

Every task must satisfy at least one requirement in PRODUCT_SPEC.md,
except an external-boundary verification task explicitly created
to resolve [UNVERIFIED] behavior required by a PRODUCT_SPEC.md requirement.

Every requirement in PRODUCT_SPEC.md that requires implementation
must be satisfied by at least one task.

TASKS.md contains only tasks whose acceptance criteria can be verified
by an automated test or an observable behavior.

Do not create tasks for content or visual design:
wording, copy, translations, images, styling, layout polish, colors.
These are human-owned and judged by a human
(see HUMAN_OWNED_CHANGES in AGENTS.md).

If PRODUCT_SPEC.md requires a content slot to exist,
the task verifies that the slot renders provided content,
not what the content says.

Each implementation task must be a vertical slice:
an observable, testable behavior from user-visible input to output.

Do not create layer-only implementation tasks.

External-boundary verification tasks are an exception to the
user-visible vertical-slice rule.
Their purpose is to establish observed behavior at an external boundary
before dependent implementation work is performed.

Each task must be completable by one AI in one session,
assuming all earlier tasks are complete.

AGENTS.md requires stopping after each task.

## Task ordering

Order tasks by dependency.

The first task is the project foundation.

Its acceptance criteria:
the constraints in TECH_STACK.md are satisfied
and the checks defined there pass.

Do not restate TECH_STACK.md.

If a PRODUCT_SPEC.md requirement depends on an external system
whose behavior is marked [UNVERIFIED], create one or more
prerequisite verification tasks before creating implementation tasks
that depend on that behavior.

Do not infer, assume, or invent the unverified behavior.

An external-boundary verification task MUST:

- identify the external boundary by reference to the relevant document section
- record the observed live behavior needed by dependent tasks
- include an executable test or probe against the live boundary
- define the observed result in pass/fail terms

After the required behavior has been verified,
dependent implementation tasks may rely on that recorded observation.

Do not create speculative implementation tasks for behavior
that has not been verified.

## External boundaries

If a task touches an external boundary as defined in AGENTS.md,
mark it.

For an external-boundary implementation task,
its acceptance criteria MUST include:

- the required live behavior is observed at the boundary
- an E2E test exercises the product through the live boundary
- the observed success behavior passes
- the failure behavior defined by ARCHITECTURE.md passes

For an external-boundary verification task,
its acceptance criteria MUST include
a recorded live observation and an executable probe or E2E check
against the live boundary.

Do not replace live-boundary verification with mocked behavior.

Mocks may be used in addition to live verification,
but a mock does not satisfy the external-boundary verification requirement.

## Acceptance criteria rules

Each task has its own acceptance criteria.

Each criterion must be pass/fail verifiable
by an automated test or an observable behavior.

Acceptance criteria MUST:

- reference the behavior required by the task
- include applicable failure behavior defined in ARCHITECTURE.md
- include unit tests passing where unit tests apply
- avoid subjective wording

A task is complete only when all its acceptance criteria are satisfied.

## Existing code

If the repository contains source code:

Still generate the full task list.
Never remove a task because it appears implemented.

Mark a task [x] only if you have verified every one of its
acceptance criteria by actually running the tests
or observing the behavior.

Code that appears to implement a task is not verification.

If any criterion cannot be verified, leave the task [ ].

An external-boundary implementation task may be marked [x] only
after its required live E2E test has passed.

An external-boundary verification task may be marked [x] only
after its required live observation and executable probe or E2E check
have passed.

If the repository has no source code, all implementation tasks are [ ].

## Execution model

TASKS.md is generated from the current contents of:

- PRODUCT_SPEC.md
- TECH_STACK.md
- ARCHITECTURE.md
- AGENTS.md

Do not use Git history to infer product requirements,
scope, acceptance criteria, or external behavior.

Do not infer missing requirements from existing code.

Do not infer missing external behavior from documentation alone
when the relevant section is marked [UNVERIFIED].

Generate TASKS.md only once.
After this, tasks are changed only as defined in AGENTS.md.

## Structure

Organize tasks in dependency order.

Do not create MVP/Future phases unless PRODUCT_SPEC.md itself explicitly
defines such phases.

Use this format for every task:

- [ ] T-001 <capability>
  - Satisfies: <PRODUCT_SPEC.md section>
  - External boundary: yes | no
  - Acceptance criteria:
    - <criterion>

For an external-boundary verification task,
use the same format and identify the referenced PRODUCT_SPEC.md section
whose required behavior is being verified.

Optimize for autonomous execution by AI systems.
