Decision-Grade Reasoning (DGR)

Audit-ready decision artifacts for LLM outputs — assumptions, risks, recommendation, and review gating (schema-valid JSON).

Security Scan

The skill is an instruction-only, schema-driven reasoning template that is internally consistent with its stated purpose and requests no unusual privileges or installs.

Purpose & Capability

Name/description (produce auditable, schema-valid decision artifacts) aligns with the provided files (SKILL.md, prompt.md, schema.json, examples). No binaries, env vars, or installs are requested — all consistent with an instruction-only formatting/templating skill.

Instruction Scope

Runtime instructions are narrowly scoped to producing a JSON artifact that conforms to schema.json, asking for clarifications when inputs are missing, and to avoid fabricating sources or chain-of-thought. It also directs generating UUIDs and computing a stable query hash (e.g., sha256 of the user query) — operations that are reasonable but will surface whatever user input is provided. Note: the skill will encode decision content into artifacts, so sensitive inputs become part of that artifact.

Install Mechanism

No install spec or code artifacts that execute on disk; instruction-only skill — lowest-risk install profile.

Credentials

Requires no environment variables, credentials, or config paths. The lack of requested secrets is proportionate to the described functionality.

Persistence & Privilege

always:false and normal model invocation behavior. The skill does not request persistent system presence or modify other skills. Users should, however, consider how/where generated artifacts are stored because they may contain sensitive decision data.

Assessment

This skill is instruction-only and internally consistent: it only formats reasoning into a strict JSON schema and requests no credentials or installs. Before using it, consider: (1) any sensitive or personally identifiable information you include in the question will appear verbatim in the artifact — ensure secure storage and retention policies; (2) for legal/medical/financial or safety-critical decisions rely on human experts (the skill explicitly gates review_required for high-stakes cases); (3) validate produced artifacts against schema.json in your pipeline and spot-check outputs for hallucinated facts or missing clarifications; and (4) test the skill with representative inputs to confirm the agent implements the

DGR — Decision‑Grade Reasoning (Governance Protocol)

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

Slug: dgr · Version: 1.0.4 · Modes: dgr_min / dgr_full / dgr_strict · Output: schema-valid JSON

What this skill does

DGR is a reasoning governance protocol that produces a machine‑validated, auditable artifact describing:

• the decision context,
• explicit assumptions and risks,
• a recommendation with rationale,
• and a consistency check.

This skill is designed for high‑stakes or review‑required decisions where you want traceability and structured review.

How to use

1. Ask your question — Provide a decision request or problem context
2. Pick mode: dgr_min | dgr_full | dgr_strict
3. Store JSON artifact in ticket / incident / audit log

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.

Mode Behavior

Mode	Speed	Detail Level	Clarifications	Review Required	Use Case
dgr_min	Fastest	Minimal compliant output	Only critical gaps	Risk-based	Quick decisions, low stakes
dgr_full	Moderate	Fuller decomposition + alternatives	More proactive	Balanced	Standard decision support
dgr_strict	Slower	Conservative analysis	More questioning	Default on ambiguity	High-stakes, uncertain contexts

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.

Files in this skill

• prompt.md — operational instructions
• schema.json — output schema (stub aligned to DGR spec)
• examples/*.md — example inputs and outputs
• field_guide.md — how to interpret DGR artifact fields

Quick start

1. Provide a decision request.
2. Choose a mode (dgr_min default).
3. The skill returns a JSON artifact suitable for review and storage.

Changelog

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.