work-intake（工作摄入量）

Work Intake

Overview

Every work request flows through intake. This skill determines scope, gathers requirements, and routes to the appropriate workflow.

Core principle: No work is too small to track, no work is too large to decompose.

Announce at start: "I'm using work-intake to understand and scope this request before beginning."

The Intake Flow

┌─────────────────────────────────────────────────────────────────────┐
│                        REQUEST RECEIVED                              │
└─────────────────────────────┬───────────────────────────────────────┘
                              │
                              ▼
┌─────────────────────────────────────────────────────────────────────┐
│               STEP 0: PROJECT BOARD READINESS (GATE)                │
│  Is GITHUB_PROJECT_NUM set?                                         │
│  Is project board accessible?                                       │
│  Are required fields configured?                                    │
└─────────────────────────────┬───────────────────────────────────────┘
                              │
                              ▼
┌─────────────────────────────────────────────────────────────────────┐
│                    CLARIFYING QUESTIONS                              │
│  What is the user trying to achieve?                                │
│  What does success look like?                                        │
│  What constraints exist?                                             │
└─────────────────────────────┬───────────────────────────────────────┘
                              │
                              ▼
┌─────────────────────────────────────────────────────────────────────┐
│                      SCOPE ASSESSMENT                                │
│  How much investigation is needed?                                   │
│  How many deliverables?                                              │
│  How many unknowns?                                                  │
└─────────────────────────────┬───────────────────────────────────────┘
                              │
              ┌───────────────┼───────────────┬───────────────┐
              ▼               ▼               ▼               ▼
         ┌────────┐     ┌──────────┐    ┌──────────┐    ┌──────────┐
         │TRIVIAL │     │  SMALL   │    │  LARGE   │    │ MASSIVE  │
         │        │     │          │    │          │    │          │
         │1 issue │     │1-3 issues│    │1 epic    │    │Initiative│
         │no unkn.│     │few unkn. │    │research  │    │multi-epic│
         └───┬────┘     └────┬─────┘    └────┬─────┘    └────┬─────┘
             │               │               │               │
             ▼               ▼               ▼               ▼
      issue-prerequisite  issue-prerequisite  epic-management  initiative-
      + project-board-    + decomposition    + research spikes  architecture
        enforcement       + project-board-  + project-board-  + project-board-
                           enforcement       enforcement       enforcement

Step 0: Project Board Readiness (GATE)

Before any work intake, verify project board infrastructure is ready.

This is a gate. Do not proceed to clarifying questions until this passes.

# Derive defaults from GITHUB_PROJECT if provided
if [ -z "$GITHUB_PROJECT_NUM" ] && [ -n "$GITHUB_PROJECT" ]; then
  NUM_CANDIDATE=$(echo "$GITHUB_PROJECT" | sed -E 's#.*/projects/([0-9]+).*#\1#')
  if [ -n "$NUM_CANDIDATE" ] && [ "$NUM_CANDIDATE" != "$GITHUB_PROJECT" ]; then
    export GITHUB_PROJECT_NUM="$NUM_CANDIDATE"
    echo "Derived GITHUB_PROJECT_NUM=$GITHUB_PROJECT_NUM from GITHUB_PROJECT"
  fi
fi

if [ -z "$GH_PROJECT_OWNER" ] && [ -n "$GITHUB_OWNER" ]; then
  export GH_PROJECT_OWNER="$GITHUB_OWNER"
  echo "Derived GH_PROJECT_OWNER=$GH_PROJECT_OWNER from GITHUB_OWNER"
fi

if [ -z "$GH_PROJECT_OWNER" ] && [ -n "$GITHUB_PROJECT" ]; then
  OWNER_CANDIDATE=$(echo "$GITHUB_PROJECT" | sed -E 's#https://github.com/(orgs|users)/([^/]+)/projects/[0-9]+#\2#')
  if [ -n "$OWNER_CANDIDATE" ] && [ "$OWNER_CANDIDATE" != "$GITHUB_PROJECT" ]; then
    export GH_PROJECT_OWNER="$OWNER_CANDIDATE"
    echo "Derived GH_PROJECT_OWNER=$GH_PROJECT_OWNER from GITHUB_PROJECT"
  fi
fi

if [ -z "$GH_PROJECT_OWNER" ]; then
  REMOTE_URL=$(git remote get-url origin 2>/dev/null || true)
  OWNER_CANDIDATE=$(echo "$REMOTE_URL" | sed -E 's#(git@|https://)github.com[:/]+([^/]+)/[^/]+(\.git)?#\2#')
  if [ -n "$OWNER_CANDIDATE" ] && [ "$OWNER_CANDIDATE" != "$REMOTE_URL" ]; then
    export GH_PROJECT_OWNER="$OWNER_CANDIDATE"
    echo "Derived GH_PROJECT_OWNER=$GH_PROJECT_OWNER from git remote"
  fi
fi

# Verify environment variables are set
if [ -z "$GITHUB_PROJECT_NUM" ]; then
  echo "BLOCKED: GITHUB_PROJECT_NUM not set"
  echo "Set with: export GITHUB_PROJECT_NUM=<number>"
  exit 1
fi

if [ -z "$GH_PROJECT_OWNER" ]; then
  echo "BLOCKED: GH_PROJECT_OWNER not set"
  echo "Set with: export GH_PROJECT_OWNER=@me  # or org name"
  exit 1
fi

# Verify project is accessible
if ! gh project view "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER" --format json > /dev/null 2>&1; then
  echo "BLOCKED: Cannot access project $GITHUB_PROJECT_NUM"
  echo "Verify project exists and you have access"
  exit 1
fi

# Verify required fields exist
FIELDS=$(gh project field-list "$GITHUB_PROJECT_NUM" --owner "$GH_PROJECT_OWNER" --format json | jq -r '.fields[].name')

for required in "Status" "Priority"; do
  if ! echo "$FIELDS" | grep -q "^$required$"; then
    echo "WARNING: Required field '$required' not found in project"
    echo "Consider adding this field for full tracking support"
  fi
done
if ! echo "$FIELDS" | grep -q "^Type$" && ! echo "$FIELDS" | grep -q "^Issue Type$"; then
  echo "WARNING: Required field 'Type' (or 'Issue Type') not found in project"
  echo "Consider adding this field for full tracking support"
fi

echo "Project board ready: $GITHUB_PROJECT_NUM"

If gate fails:

1. Configure missing environment variables
2. Create project board if needed
3. Add required fields (Status, Type, Priority)
4. Re-run readiness check

Skill: project-board-enforcement

Step 1: Clarifying Questions

Before scoping, understand the request.

Do not ask questions that can be answered from the repo. First inspect:

• README.md, FEATURES.md, BRANDING.md, docs/, Storybook, and existing routes/pages
• Existing GitHub issues and project board items

Only ask the user for information that is still missing after reviewing the repo.

Essential Questions

Discovery Questions (for unclear requests)

Step 2: Scope Assessment

Evaluate the request against these criteria:

Scope Matrix

Scope Decision

IF unknowns == none AND deliverables <= 2:
  → TRIVIAL: Use issue-prerequisite directly

IF unknowns == few AND deliverables <= 5:
  → SMALL: Use issue-prerequisite, maybe issue-decomposition

IF unknowns == many OR deliverables > 5:
  → LARGE: Use epic-management with research spikes

IF unknowns == extensive OR deliverables > 20 OR new_infrastructure:
  → MASSIVE: Use initiative-architecture

Step 3: Route to Appropriate Workflow

Trivial Scope

**Assessment:** This is a trivial request (single deliverable, no unknowns).

**Next step:** Creating a single issue using `issue-prerequisite`.

Route to: issue-prerequisite

Small Scope

**Assessment:** This is a small request (few deliverables, minimal unknowns).

**Plan:**
1. Create parent issue for the request
2. If needed, decompose into 2-3 sub-issues
3. Begin implementation

**Next step:** Creating issue structure using `issue-prerequisite`.

Route to: issue-prerequisite → maybe issue-decomposition

Large Scope

**Assessment:** This is a large request requiring structured planning.

**Unknowns identified:**
- [ ] Unknown 1 - needs investigation
- [ ] Unknown 2 - needs research spike

**High-level deliverables:**
1. Deliverable A
2. Deliverable B
...

**Next step:** Creating epic structure using `epic-management`.

Route to: epic-management (which will create research spikes as needed)

Massive Scope

**Assessment:** This is a massive request requiring full initiative architecture.

**Why massive:**
- [Reason: extensive unknowns / new infrastructure / multi-system / etc.]

**Initial unknowns:**
- [ ] Does X exist?
- [ ] How does Y work?
- [ ] What are constraints of Z?

**Potential scope:**
- Multiple epics likely
- New capabilities needed
- Significant research required

**Next step:** Beginning initiative architecture using `initiative-architecture`.

Route to: initiative-architecture

Resumability Checkpoint

Before routing, document the intake in memory:

# Store intake assessment
mcp__memory__create_entities([{
  "name": "Intake-[DATE]-[SHORT_DESC]",
  "entityType": "WorkIntake",
  "observations": [
    "Request: [Original request]",
    "Scope: [Trivial/Small/Large/Massive]",
    "Unknowns: [List]",
    "Route: [Target skill]",
    "Status: [Routing/In Progress/Complete]"
  ]
}])

Examples

Example: Trivial

Request: "Make the login page button a little lighter."

Intake:

• Goal: Adjust button color
• Success: Button is lighter
• Unknowns: None (CSS change)
• Deliverables: 1 (color change)
• Scope: TRIVIAL

→ Route to issue-prerequisite

Example: Large

Request: "Add dark mode to the application."

Intake:

• Goal: Theme switching capability
• Success: Users can toggle dark/light mode
• Unknowns:

- Current theming approach? - Component library support? - Persistence mechanism?

• Deliverables: ~10 (theme tokens, components, toggle, persistence, etc.)
• Scope: LARGE

→ Route to epic-management

Example: Massive

Request: "Add the ability for users to log in by clicking on a popup in their phone."

Intake:

• Goal: Mobile push notification login
• Success: User receives push, taps, logged in
• Unknowns:

- Is there a mobile app? → Research spike - Push notification infrastructure? → Research spike - Authentication flow design? → Research spike - Security requirements? → Research spike

• Deliverables: 30+ (app work, backend, auth, notifications, etc.)
• Scope: MASSIVE

→ Route to initiative-architecture

Red Flags That Indicate Larger Scope

Watch for these phrases that suggest the request is larger than it appears:

Never Turn Away Work

No matter how massive the request:

1. Acknowledge the goal - "I understand you want X"
2. Explain the process - "This will require structured planning"
3. Begin investigation - Start with what we don't know
4. Document everything - Every decision, every finding
5. Break it down - Until we have tractable pieces

The path from "massive request" to "implementation" is:

Massive Request
    → Initiative Architecture (document unknowns, create research spikes)
    → Research Spikes (answer unknowns)
    → Epic Structure (group deliverables)
    → Issue Decomposition (create tractable tasks)
    → Implementation (one issue at a time)

Checklist

• Project board readiness verified (Step 0 gate passed)
• GITHUB_PROJECT_NUM and GH_PROJECT_OWNER set
• Asked clarifying questions
• Understood the goal (not just the task)
• Identified unknowns
• Counted deliverables
• Assessed scope (Trivial/Small/Large/Massive)
• Documented intake in memory
• Routed to appropriate skill (with project-board-enforcement)

Skill: project-board-enforcement

Gate: Cannot proceed to any downstream skill without project board readiness verified.