handoff

Name: handoff
Author: anombyte93

582stars57forks

How to install handoff?

anombyte93/prd-taskmaster/handoff

$npx -y skills add anombyte93/prd-taskmaster --skill handoff

Installs into the current project.

Summary

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 w

Files · 1

View on GitHub

SKILL.md

1	# Phase 3: Handoff
2
3	Declarative phase skill. Invoked by the prd-taskmaster orchestrator when
4	`current_phase` is `HANDOFF`. Never called directly by a user.
5
6	The one rule: **detect what the user has, recommend ONE mode, give the user a
7	structured choice, dispatch the chosen mode. Mode D executes only on tier=premium; otherwise it is a locked teaser.**
8
9	## Entry gate
10
11	1. Call `mcp__plugin_prd_go__check_gate(phase="HANDOFF", evidence={})` for diagnostics.
12
13	`check_gate` is an EXIT gate: it requires `user_mode_choice` and `plan_file_exists` —
14	both produced by HANDOFF itself, i.e. evidence to advance, not to enter. On first
15	entry neither exists yet, so a `gate_passed: false` here is EXPECTED — the state
16	machine's legal transitions already guarantee only legal entry.
17
18	- First entry (no evidence yet): note the result and continue with the Procedure.
19	- Re-entry: if the gate reports violations, report them and stop — it protects
20	against re-running a completed phase or skipping ahead from GENERATE.
21	2. Read the GENERATE outputs — `.taskmaster/docs/prd.md`, `.taskmaster/tasks/tasks.json`,
22	`.taskmaster/reports/task-complexity-report.json`. If any are missing,
23	report and stop. The gate should have caught this, but belt-and-braces.
24
25	## Handoff checklist
26
27	Copy into your response before running the procedure:
28
29	```
30	HANDOFF CHECKLIST:
31	- [ ] Capabilities detected (tier + per-capability flags)
32	- [ ] Recommended mode: ___ (reason: ___)
33	- [ ] Summary displayed (spec location, task count, capabilities)
34	- [ ] CLAUDE.md task workflow appended (idempotent)
35	- [ ] AskUserQuestion mode picker surfaced (or prose fallback if hook-blocked)
36	- [ ] User choice dispatched (Mode A / B / C, or D when tier=premium)
37	- [ ] Debrief scaffold emitted (optional, silently tolerated)
38	- [ ] Handoff complete
39	```
40
41	## Step 1: Detect capabilities
42
43	MCP (preferred): `mcp__plugin_prd_go__detect_capabilities()`
44
45	CLI fallback: `python3 script.py detect-capabilities`
46
47	Returns a `tier` field (`"free"` or `"premium"`) plus per-capability flags.
48	Key signals:
49
50	\| Capability \| What It Enables \|
51	\|------------\|----------------\|
52	\| superpowers plugin \| Modes A, C (brainstorm, plans, subagents) \|
53	\| task-master-ai (CLI or MCP) \| Mode B (native auto-execute loop) \|
54	\| ralph-loop plugin \| Mode C (iterative execution loop) \|
55	\| atlas-launcher MCP (licensed) \| Mode D — Atlas Fleet (tier=premium) \|
56	\| atlas-loop / atlas-cdd skills \| legacy Mode-D seeds — superseded by atlas-launcher detection \|
57	\| Research model (task-master or MCP) \| Deep research per task \|
58	\| Playwright MCP \| Tier S browser verification \|
59
60	Mode D (Atlas Fleet) unlocks on `tier: "premium"` only — i.e. a licensed
61	`atlas-launcher` MCP registration detected by `detect_atlas_launcher()`. Local
62	`atlas-loop`/`atlas-cdd` skills do NOT unlock it. See Step 2 and the Mode D
63	section below.
64
65	## Step 2: Recommend ONE mode
66
67	Decision logic (first match wins):
68
69	- `superpowers` + `ralph-loop` present → Mode C (recommended free)
70	- `superpowers` only → Mode A (plan-only, manual drive)
71	- `task-master-ai` only → Mode B (native auto-execute)
72	- Fallback → Mode A
73
74	External-tool modes (E–J: Cursor, RooCode, Codex, Gemini, CodeRabbit, Aider)
75	are offered as alternatives via the `alternative_modes` field, not primary
76	recommendations. **Mode D is recommended iff `tier == "premium"` AND the task
77	graph parallelizes (>= 2 independent dependency chains — check `fleet-waves`
78	output: any wave with >= 2 chunks). Premium + serial graph: recommend the best
79	free mode and say why ("your tasks form a single dependency chain — Verified
80	Loop is the right tool here"); Fleet stays selectable but not default. Free
81	tier: Mode D is a locked Atlas Pro teaser, never selectable, regardless of
82	which local plugins are installed.**
83
84	### Mode A: Plan Only (Manual)
85
86	```
87	Recommended: Plan Only
88	superpowers:writing-plans creates your implementation plan
89	Plan references TaskMaster task IDs from tasks.json
90	You drive execution manually
91	```
92
93	### Mode B

People who liked this also viewed

6 picks

Type

handoff

582stars57forks

View on GitHub

How to install handoff?

anombyte93/prd-taskmaster/handoff

$npx -y skills add anombyte93/prd-taskmaster --skill handoff

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Phase 3: Handoff
2
3	Declarative phase skill. Invoked by the prd-taskmaster orchestrator when
4	`current_phase` is `HANDOFF`. Never called directly by a user.
5
6	The one rule: **detect what the user has, recommend ONE mode, give the user a
7	structured choice, dispatch the chosen mode. Mode D executes only on tier=premium; otherwise it is a locked teaser.**
8
9	## Entry gate
10
11	1. Call `mcp__plugin_prd_go__check_gate(phase="HANDOFF", evidence={})` for diagnostics.
12
13	`check_gate` is an EXIT gate: it requires `user_mode_choice` and `plan_file_exists` —
14	both produced by HANDOFF itself, i.e. evidence to advance, not to enter. On first
15	entry neither exists yet, so a `gate_passed: false` here is EXPECTED — the state
16	machine's legal transitions already guarantee only legal entry.
17
18	- First entry (no evidence yet): note the result and continue with the Procedure.
19	- Re-entry: if the gate reports violations, report them and stop — it protects
20	against re-running a completed phase or skipping ahead from GENERATE.
21	2. Read the GENERATE outputs — `.taskmaster/docs/prd.md`, `.taskmaster/tasks/tasks.json`,
22	`.taskmaster/reports/task-complexity-report.json`. If any are missing,
23	report and stop. The gate should have caught this, but belt-and-braces.
24
25	## Handoff checklist
26
27	Copy into your response before running the procedure:
28
29	```
30	HANDOFF CHECKLIST:
31	- [ ] Capabilities detected (tier + per-capability flags)
32	- [ ] Recommended mode: ___ (reason: ___)
33	- [ ] Summary displayed (spec location, task count, capabilities)
34	- [ ] CLAUDE.md task workflow appended (idempotent)
35	- [ ] AskUserQuestion mode picker surfaced (or prose fallback if hook-blocked)
36	- [ ] User choice dispatched (Mode A / B / C, or D when tier=premium)
37	- [ ] Debrief scaffold emitted (optional, silently tolerated)
38	- [ ] Handoff complete
39	```
40
41	## Step 1: Detect capabilities
42
43	MCP (preferred): `mcp__plugin_prd_go__detect_capabilities()`
44
45	CLI fallback: `python3 script.py detect-capabilities`
46
47	Returns a `tier` field (`"free"` or `"premium"`) plus per-capability flags.
48	Key signals:
49
50	\| Capability \| What It Enables \|
51	\|------------\|----------------\|
52	\| superpowers plugin \| Modes A, C (brainstorm, plans, subagents) \|
53	\| task-master-ai (CLI or MCP) \| Mode B (native auto-execute loop) \|
54	\| ralph-loop plugin \| Mode C (iterative execution loop) \|
55	\| atlas-launcher MCP (licensed) \| Mode D — Atlas Fleet (tier=premium) \|
56	\| atlas-loop / atlas-cdd skills \| legacy Mode-D seeds — superseded by atlas-launcher detection \|
57	\| Research model (task-master or MCP) \| Deep research per task \|
58	\| Playwright MCP \| Tier S browser verification \|
59
60	Mode D (Atlas Fleet) unlocks on `tier: "premium"` only — i.e. a licensed
61	`atlas-launcher` MCP registration detected by `detect_atlas_launcher()`. Local
62	`atlas-loop`/`atlas-cdd` skills do NOT unlock it. See Step 2 and the Mode D
63	section below.
64
65	## Step 2: Recommend ONE mode
66
67	Decision logic (first match wins):
68
69	- `superpowers` + `ralph-loop` present → Mode C (recommended free)
70	- `superpowers` only → Mode A (plan-only, manual drive)
71	- `task-master-ai` only → Mode B (native auto-execute)
72	- Fallback → Mode A
73
74	External-tool modes (E–J: Cursor, RooCode, Codex, Gemini, CodeRabbit, Aider)
75	are offered as alternatives via the `alternative_modes` field, not primary
76	recommendations. **Mode D is recommended iff `tier == "premium"` AND the task
77	graph parallelizes (>= 2 independent dependency chains — check `fleet-waves`
78	output: any wave with >= 2 chunks). Premium + serial graph: recommend the best
79	free mode and say why ("your tasks form a single dependency chain — Verified
80	Loop is the right tool here"); Fleet stays selectable but not default. Free
81	tier: Mode D is a locked Atlas Pro teaser, never selectable, regardless of
82	which local plugins are installed.**
83
84	### Mode A: Plan Only (Manual)
85
86	```
87	Recommended: Plan Only
88	superpowers:writing-plans creates your implementation plan
89	Plan references TaskMaster task IDs from tasks.json
90	You drive execution manually
91	```
92
93	### Mode B

People who liked this also viewed

6 picks

Type