From ccds-ai
Prompt design specialist. Auto-invoked when prompts, system messages, few-shot examples, structured-output schemas, or chain-of-prompt chains are being written or iterated.
How this skill is triggered — by the user, by Claude, or both
Slash command
/ccds-ai:ai-prompt-engineerThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Vague prompts are how LLM apps regress silently. A prompt is code: specific,
Vague prompts are how LLM apps regress silently. A prompt is code: specific, versioned, schema-bound, and covered by a regression run.
results array") instead of hoping.[system]
Role and task in 2–3 sentences. Hard constraints.
Output schema or format spec (the same one the code validates against).
Edge-case behavior: empty input, no answer found, conflicting sources.
[examples] # optional, 1–3, format-matched to current schema
input → expected output
[context] # dynamic, delimited, declared as data
<documents>
<doc id="1" source="...">…retrieved chunk…</doc>
</documents>
[user]
The actual input — then a one-line restatement of the task.
Related: ai-rag (the context being injected), ai-eval (regression runs on
prompt changes), ai-safety (injection defense for untrusted context),
api-design (tool/output schemas as API contracts) · domain agent:
ai-architect (model selection) · output/ADR format: playbook-conventions
npx claudepluginhub ggrace519/claude-code-dev-studio --plugin ccds-aiProvides UI/UX resources: 50+ styles, color palettes, font pairings, guidelines, charts for web/mobile across React, Next.js, Vue, Svelte, Tailwind, React Native, Flutter. Aids planning, building, reviewing interfaces.
Fetches up-to-date documentation from Context7 for libraries and frameworks like React, Next.js, Prisma. Use for setup questions, API references, and code examples.