Decision-Grade Reasoning (DGR)

Audit-ready decision artifacts for LLM outputs - assumptions, risks, recommendation, and review gating (schema-valid JSON). Replaced with https://clawhub.ai/dgr-ai-labs/plugins/openclaw-dgr-gate)

Superseded — kept as a reference. This skill formats reasoning into a structured record. It cannot stop an action; the model decides whether to follow it.

A rule in the prompt is a request. A rule in the tool is a gate. The enforcing successor is DGR Gate, a pre-execution gate that checks a call against operator-configured policy before it runs. It currently gates two demo tools, not your real ones.

This page remains published as the record of the earlier approach and why it was replaced. Existing installs keep working; no further development.

DGR — Decision‑Grade Reasoning (Governance Protocol)

Purpose: produce an auditable, machine‑validated decision record for review and storage.

Slug: dgr · Version: 1.1.0 · Modes: dgr_min / dgr_full / dgr_strict · Output: schema-valid JSON

What this skill does

This skill produced a structured decision record: decision context, explicit assumptions, identified risks, a recommendation with rationale, and a consistency check. The output was documentation of reasoning, formatted as schema-valid JSON for easier review and storage.

That was insufficient. A model's own justification is not evidence — a better-formatted rationale from the same model does not change that. A prompt-level instruction can be lost during context compaction or ignored after the model reads untrusted text. This approach documented reasoning but could not enforce constraints on tool execution.

What this skill is NOT (non‑claims)

This skill does NOT guarantee:

• correctness, optimality, or truth,
• elimination of hallucinations,
• legal/medical/financial advice suitability,
• or regulatory compliance by itself.

DGR improves process quality (clarity, traceability, reviewability) — not outcome certainty.

When to use

Use when you need:

• an auditable record of reasoning,
• explicit assumptions/risks surfaced,
• reviewer‑friendly structure,
• a consistent output format across tasks and models.

Inputs

• A user request/question (free text).
• Optional: context identifiers (ticket ID, policy name), and desired mode: dgr_min, dgr_full, or dgr_strict.

Outputs

A single JSON artifact matching schema.json.

Minimum acceptance criteria (see schema.json):

• at least 1 assumption
• at least 1 risk
• recommendation present
• consistency_check present

Safety / governance boundaries

• Always ask for clarification if key decision inputs are missing.
• If the decision is high‑risk, escalate via recommendation.review_required = true.
• If uncertainty is high, explicitly state uncertainty and limit scope.
• Do not fabricate sources or cite documents you did not see.

Changelog

1.1.0 — Repositioned as a superseded reference. Terminology aligned with DGR Gate (pre-execution gate). Removed usage guidance; this skill formats reasoning and does not enforce. Existing installs unaffected.

1.0.4 — Remove redundant CLAWHUB_SUMMARY.md; summary now sourced from SKILL.md front-matter.

1.0.3 — Tighten front-matter description for better conversion, add reasoning category, compress identity block for faster scanning.

1.0.2 — Add ClawHub front-matter metadata with emoji and homepage for improved discovery and presentation.

1.0.0 — Initial public release of DGR skill bundle with auditable decision reasoning framework, governance protocols, and structured output format.

Note: This is an opt‑in reasoning mode. It is meant to be used alongside human decision‑making, not as a replacement.