AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Next

skill-endlessblink-master-plan-next · by endlessblink

Analyze and pick the next task to work on. Reads MASTER_PLAN.md, scores tasks by priority and status, and presents interactive selection. Use when starting a session or deciding what to tackle next.

No reviews yet
0 installs
38 views
0.0% view→install

Install

$ agentstack add skill-endlessblink-master-plan-next

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-endlessblink-master-plan-next)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
6mo ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Next? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

What's Next?

Analyze MASTER_PLAN.md tasks, score by priority/status, and let the user pick interactively.

Triggers

  • /master-plan:next - Main command
  • "what should I work on", "pick a task", "next task", "what's next"

Arguments

| Argument | Filter | |----------|--------| | bugs | Only BUG-XXX tasks | | progress | Only IN PROGRESS tasks | | planned | Only PLANNED (backlog) tasks | | review | Only REVIEW tasks | | active | IN PROGRESS + REVIEW | | all | Include DONE tasks |

Example: /master-plan:next bugs or /master-plan:next planned

Workflow

Step 1: Find MASTER_PLAN.md

Search for the plan file in order:

  1. docs/MASTER_PLAN.md
  2. MASTER_PLAN.md
  3. master-plan.md
  4. docs/master-plan.md

If not found, tell the user: "No MASTER_PLAN.md found. Run /master-plan:task to create your first task, or create one from the template."

Step 2: Parse Tasks

Read the file and extract tasks from ### headers matching this pattern:

### [optional ~~]TASK-123[optional ~~]: Title (STATUS)

Regex: ^###\s+(~~)?((TASK|BUG|FEATURE|ROAD|IDEA|ISSUE|INQUIRY)-\d+)(~~)?:\s*(.+?)\s*\(([^)]+)\)

For each match, extract:

  • ID: e.g., TASK-123, BUG-456
  • Title: The text after : and before (
  • Status: Text in parentheses. Normalize:
  • Contains "DONE", "COMPLETE", "FIXED", "✅", or has strikethrough → DONE
  • Contains "IN PROGRESS", "🔄" → IN PROGRESS
  • Contains "REVIEW", "👀" → REVIEW
  • Contains "PAUSED", "⏸️" → PAUSED
  • Otherwise → PLANNED
  • Priority: Look in the lines following the header for **Priority**: P0-P3. Default to P2 if not found.

Step 3: Check for Active Work

Look for IN PROGRESS tasks first — these should be finished before starting new work.

Also check git status:

git status --short

If uncommitted changes exist, mention: "You have uncommitted changes — consider committing or running /master-plan:save first."

Step 4: Sort and Filter

Default sort order (what to START next):

  1. PLANNED first, then REVIEW, IN PROGRESS, PAUSED, DONE last
  2. Within same status: P0 → P1 → P2 → P3

If $ARGUMENTS contains a filter (bugs, progress, planned, review, active), apply it before sorting.

Exclude DONE tasks unless $ARGUMENTS contains all.

Step 5: Present Interactive Selection

Use AskUserQuestion to let user pick from the top tasks:

AskUserQuestion({
  questions: [{
    question: "Which task would you like to work on?",
    header: "Task",
    multiSelect: false,
    options: [
      // Top 4 tasks, formatted as:
      { label: "TASK-XXX: Title here", description: "P0 · IN PROGRESS" },
      { label: "BUG-YYY: Another task", description: "P1 · PLANNED" },
      // ...
    ]
  }]
})

Option formatting:

  • Label: {ID}: {title (max 40 chars)}
  • Description: {priority} · {status}

If there are IN PROGRESS tasks, always show them first with a note: "You have N task(s) in progress."

Step 6: Show Task Details

When user selects a task, show the full section from MASTER_PLAN.md:

  • ID, title, status, priority
  • Description (everything between this ### header and the next ### or ##)

Step 7: Offer Actions

After showing task details, ask:

AskUserQuestion({
  questions: [{
    question: "What would you like to do?",
    header: "Action",
    multiSelect: false,
    options: [
      { label: "Start working on this", description: "Update status to IN PROGRESS" },
      { label: "Pick a different task", description: "Go back to task list" },
      { label: "Just show context", description: "No status change" }
    ]
  }]
})

If "Start working on this" is selected:

  1. Update the task's status to IN PROGRESS in MASTER_PLAN.md (all locations — see update rules below)
  2. Confirm: "Started TASK-XXX. Ready to begin implementation."

MASTER_PLAN.md Update Rules

Tasks may appear in multiple locations. When updating status, check ALL of:

  1. Summary/Roadmap table — Update the status column
  2. Detailed ### section header — Update (STATUS) in parentheses
  3. Subtask bullet points — Add strikethrough + ✅ if marking DONE

Always verify with a grep after updating.

Rules

  1. Finish before starting — Always highlight IN PROGRESS tasks first
  2. P0 trumps all — Critical issues come first regardless of status
  3. Interactive selection — Always use AskUserQuestion, never just print a list
  4. Context on selection — Always show full task details when picked
  5. Action oriented — Offer to start work immediately

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.