Agent skill · Testing & QA

handoff

Phase 3 of the prd-taskmaster pipeline: smart mode selection and user handoff. Detects installed capabilities (superpowers, ralph-loop, task-master-ai, playwright, research providers), recommends ONE execution mode (A/B/C) with reasoned justification, appends the task-execution workflow to CLAUDE.md, surfaces a structured AskUserQuestion multi-option picker for user agency, and dispatches the chosen mode. Mode D (Atlas Fleet) is selectable only when detect_capabilities returns tier=premium (licensed atlas-launcher detected); otherwise it is a locked Atlas Pro teaser. Plan Mode is NOT used (spe

anombyte93github.com/anombyte93GitHub ↗
claude-coderead-onlyMIT
Install
npx skills add anombyte93/prd-taskmaster --skill handoff --agent claude-code

Same command for any agent — swap --agent for codex, cursor, copilot.

Facts
Files in the skill folder: 1
SKILL.md size: 20 KB
Bundled scripts: none
Allowed tools: -Read-Skill-AskUserQuestion-ToolSearch-mcp__atlas-engine-mcp__plugin_prd_go
Path: skills/handoff/SKILL.md
Open the folder on GitHub →
Where it comes from
Stars: 586
Language: Python

Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.

Review
written from the skill's own SKILL.md · Aug 5, 2026

What it does

Phase 3 of the prd-taskmaster pipeline: smart mode selection and user handoff. Detects installed capabilities, recommends ONE execution mode with reasoned justification, appends the task-execution workflow to CLAUDE.md, surfaces a structured AskUserQuestion multi-option picker for user agency, and dispatches the chosen mode. Mode D (Atlas Fleet) is selectable only when detect_capabilities returns tier=premium; otherwise it is a locked Atlas Pro teaser. Plan Mode is NOT used and AskUserQuestion is the sole user-agency mechanism. Declares HANDOFF complete so EXECUTE can follow.

How it works

  • Entry gate: runs diagnostics via mcp__plugin_prd_go__check_gate(phase="HANDOFF", evidence={}) and then reads outputs from GENERATE artifacts. It ensures required evidence (e.g., user_mode_choice, plan_file_exists) is produced by HANDOFF for advancement.
  • Step 1: Detect capabilities using MCP preferred function mcp__plugin_prd_go__detect_capabilities() (or CLI fallback) to obtain a tier value ("free" or "premium") and per-capability flags.
  • Step 2: Recommend ONE mode using first-match rules: if superpowers and ralph-loop present → Mode C; if superpowers only → Mode A; if task-master-ai only → Mode B; otherwise Mode A. Mode D unlocks only when tier == "premium" and graph parallelizes. Alternatives E–J are offered via alternative_modes but not as default.
  • Step 3: Append the task workflow to CLAUDE.md using a deterministic, idempotent subcommand path that writes a fixed workflow snippet, with safeguards and HTML-comment sentinels.
  • Step 4: Display a summary panel with validation and capabilities, then show a handoff surface including mode options via AskUserQuestion.
  • Step 5: Mandatory AskUserQuestion for mode selection with options: Plan & Drive (Mode A), Auto-Execute (Mode B), Verified Loop (Mode C), Atlas Fleet (Mode D), Use another tool… (E–J), Show me more detail before I decide. The recommended mode is set as default when applicable (Atlas Fleet on premium with graph parallelization).
  • Dispatch: depending on the chosen mode, perform corresponding actions:
    • Mode A: call superpowers:writing-plans with spec path .taskmaster/docs/prd.md.
    • Mode B: show commands for next task (CLI or MCP backend) and proceed with native TaskMaster execution loop.
    • Mode C: write .claude/atlas-loop-prompt.md and invoke /goal with the specified SHIP_CHECK_OK/conditions until completion.
    • Mode D: invoke /prd:execute-fleet (premium) or present upgrade message on free tier.

When to use it

Use when the orchestrator is in HANDOFF phase and needs to determine a single, justified execution mode, present user agency, and dispatch the selected mode. Mode D is only selectable on premium installations.

What it can touch

  • Tools: Read, Skill, AskUserQuestion, ToolSearch, mcp__atlas-engine, mcp__plugin_prd_go, mcp__plugin_prd-taskmaster_go, mcp__plugin_atlas-go_go
  • Files: CLAUDE.md (workflow appended), .taskmaster/docs/prd.md, .taskmaster/tasks/tasks.json, .taskmaster/reports/task-complexity-report.json, .claude/atlas-loop-prompt.md

Caveats

  • Mode D is locked on free tier; attempting to select it prompts an upgrade message and re-prompts with free modes.
  • Plan Mode is explicitly NOT used.
  • Requires generation outputs to exist; otherwise it reports and stops per gate behavior.
  • The skill is user-invocable false; it is invoked by the prd-taskmaster orchestrator when phase HANDOFF is active.
From the SKILL.md

# Phase 3: Handoff Declarative phase skill. Invoked by the prd-taskmaster orchestrator when `current_phase` is `HANDOFF`. Never called directly by a user. The one rule: **detect what the user has, recommend ONE mode, give the user a structured choice, dispatch the chosen mode. Mode D executes only on tier=premium; otherwise it is a locked teaser.** ## Entry gate 1. Call `mcp__plugin_prd_go__check_gate(phase="HANDOFF", evidence={})` for diagnostics. `check_gate` is an EXIT gate: it requires `user_mode_choice` and `plan_file_exists` — both produced by HANDOFF itself, i.e. evidence to *advance*, not to *enter*. On first entry neither exists yet, so a `gate_passed: false` here is EXPECTED — the state machine's legal transitions already guarantee only legal entry. - **First entry** (no evidence yet): note the result and continue with the Procedure. - **Re-entry**: if the gate reports violations, report them and stop — it protects against re-running a completed phase or skipping ahead from GENERATE. 2. Read the GENERATE outputs — `.taskmaster/docs/prd.md`, `.taskmaster/tasks/tasks.json`, `.taskmaster/reports/task-complexity-report.json`. If any are missing, report and stop. The gate shou

What's inside
Steps it walks through
  1. Entry gate
  2. Handoff checklist
  3. Step 1: Detect capabilities
  4. Step 2: Recommend ONE mode
  5. Mode A: Plan Only (Manual)
  6. Mode B: TaskMaster Auto-Execute (Mode B — TaskMaster backend only)
  7. Mode C: Plan + Ralph Loop (Recommended Free)
  8. Mode D: Atlas Fleet (selectable on tier=premium; 🔒 locked teaser on free)
  9. Alternative modes E–J (external AI tools)
  10. Step 3: Append task workflow to CLAUDE.md
  11. Step 4: Display summary
  12. Step 5: Mandatory AskUserQuestion for mode selection
  13. Sequence
  14. Hook-blocked fallback (graceful degradation)
Commands it runs
python3 $SKILL_DIR/script.py append-workflow \
python3 "$SKILL_DIR/script.py" debrief \
More from prd-taskmaster
All skills →
About this skill
What does the handoff skill do?

Phase 3 of the prd-taskmaster pipeline: smart mode selection and user handoff. Detects installed capabilities (superpowers, ralph-loop, task-master-ai, playwright, research providers), recommends ONE execution mode (A/B/C) with reasoned justification, appends the task-execution workflow to CLAUDE.md, surfaces a structured AskUserQuestion multi-option picker for user agency, and dispatches the chosen mode. Mode D (Atlas Fleet) is selectable only when detect_capabilities returns tier=premium (licensed atlas-launcher detected); otherwise it is a locked Atlas Pro teaser. Plan Mode is NOT used (spe

How do I install it?

Run `npx skills add anombyte93/prd-taskmaster --skill handoff --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.

Where does this skill come from?

From anombyte93/prd-taskmaster, a repository with 586 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.

Is a popular skill a good skill?

Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.

Keep going