Prompting Frameworks — Testing skill for Claude Code

Prompting Frameworks

Four Claude/Cowork skills for structuring and QA'ing prompts, built around GRC and audit-facing work but generally applicable. Each `.skill` file is a ready-to-install package (zip a folder containing `SKILL.md` — the standard Claude Code / Cowork skill format); import it directly, or read the `SKILL.md` inside for the technique on its own.

See `AI_Prompting_Frameworks_Reference_Guide.docx` for the full writeup tying all four together.

The four skills

Skill	Use it for
art-framework-prompting.skill	Refining a vague, underspecified one-line ask before acting on it — a personal convention, not an industry-standard framework
icio-framework-prompting.skill	Structuring production LLM/API prompts (Instruction / Context / Input data / Output format) for pipelines where parseable output matters — n8n steps, scoring calls, structured DB writes
cot-framework-prompting.skill	Chain-of-Thought reasoning for multi-factor judgment calls (risk tiering, control effectiveness ratings, severity classification) before committing to a verdict — this one is a real, citable industry-standard technique, safe to reference by name in client-facing material
reflexion-self-critique.skill	A single-pass draft → critique-against-rubric → revise loop for any deliverable before finalizing it — audit findings, client memos, training content

**ART vs ICIO vs CoT vs Reflexion, in one line each:**

• ART = sharpening what you ask Claude, conversationally
• ICIO = structuring a prompt that runs unattended against an API/parser
• CoT = forcing explicit reasoning before a judgment call
• Reflexion = QA'ing a finished draft against an explicit rubric before shipping it

ART and ICIO are personal conventions (not external frameworks — don't cite them as industry standard in client-facing material). CoT is a real, citable technique. Reflexion is a lightweight, single-model variant of self-critique prompting.

Install

Each `.