rskill-playbook-decompose_mission

rskill-playbook-decompose_mission

A kind: playbook rSkill: a symbolic S2 decision procedure the Reasoner reads, not a neural policy. It carries no weights — the authored PLAYBOOK.md is its runtime.

What this skill does

Breaks a compound, multi-step instruction into an ordered list of subtasks, each with its own verifiable done-condition (an internal TODO list), then executes and verifies them in order. It decomposes the goal, records the subtasks to memory so the plan survives a tick, dispatches the matching skill for each, verifies the done-condition before advancing, and on a subtask failure replans that subtask only — escalating to a human if a subtask exhausts its replan budget. Concrete walkthrough: the stack-bowls / drawer / cookie-box example in PLAYBOOK.md.

How it works

This playbook is content, not code. When installed, the reasoner injects PLAYBOOK.md into its system prompt and follows the SOP, composing tools it already has (execute_rskill, query_scene, query_task_progress, memory_write, memory_search, emit_prompt). It is role: s2 and is never dispatched through ExecuteSkill. Every motion it triggers is an execute_rskill → Action chunk → C++ safety kernel — the playbook holds no actuation authority (CLAUDE.md §1.1).

Observation → action contract

None. A playbook emits no Action chunks and requires no actuators (actuators_required: [], chunk_size: 1). Its "output" is the sequence of tool calls the reasoner makes while following the SOP, bounded by playbook.max_steps.

How it was authored / Upstream provenance

N/A — a playbook is hand-authored, not trained: it has no weights and no upstream model. Its provenance is the authoring decision record (also linked via paper_url). To change behaviour, edit PLAYBOOK.md and bump version.

Supported robots

Embodiment-agnostic — declares the explicit wildcard embodiment_tags: ["any"] (never an empty list). Pure planning / orchestration, so capabilities_required is empty ({}): it works on any robot. Each subtask it dispatches is gated at runtime by the composed tools, not by this playbook's flags.

Sensors required

None directly. The tools it composes declare their own sensor needs.

Manifest summary

• kind: playbook, role: s2, actions: [plan], chunk_size: 1.
• playbook.trigger: the goal is a compound, multi-step instruction.
• playbook.done_predicate: every subtask's verifiable goal has been confirmed met, or the mission has been handed off.
• playbook.max_steps: 24.

Quick start

from openral_core.schemas import RSkillManifest

m = RSkillManifest.from_yaml("rskills/decompose-mission/rskill.yaml")
assert m.kind == "playbook" and m.playbook is not None
print(m.playbook.trigger)

Reproduction

Packaging-only: the manifest + SOP are validated by tests/unit/test_playbook_rskill_manifest.py. There is no benchmark number to reproduce; the playbook's behaviour is exercised by the reasoner integration tests in later phases.

Evaluation

N/A — no eval/*.json; a playbook produces no benchmarkable policy output.

License

• Code / content: Apache-2.0.
• Weights: none.

See also

• PLAYBOOK.md — the decision procedure itself.