AgentSkillsCN

perplexity-tools

通过统一的 Python CLI 执行 Perplexity AI 查询。当您需要进行网络调研、提出带引用的问题、开展深度研究、解决推理难题，或寻找最新资讯时，可使用此技能。

SKILL.md
--- frontmatter
name: perplexity-tools
type: complex
depth: base
description: >-
  Executes Perplexity AI queries via unified Python CLI. Use when conducting
  web research, asking questions with citations, deep research tasks, reasoning
  problems, or searching for up-to-date information.

[H1][PERPLEXITY-TOOLS]

Dictum: Specialized models optimize response quality.

<br>

Execute Perplexity AI queries via unified Python CLI.

[IMPORTANT] Commands require --query. API key auto-injected via 1Password.

bash
# Quick question—citations (sonar)
uv run .claude/skills/perplexity-tools/scripts/perplexity.py ask --query "What is Effect-TS?"

# Deep research (sonar-deep-research)
uv run .claude/skills/perplexity-tools/scripts/perplexity.py research --query "React 19 new features"
uv run .claude/skills/perplexity-tools/scripts/perplexity.py research --query "Vite 7 migration" --strip-thinking

# Reasoning task (sonar-reasoning-pro)
uv run .claude/skills/perplexity-tools/scripts/perplexity.py reason --query "Compare Vite vs Webpack"
uv run .claude/skills/perplexity-tools/scripts/perplexity.py reason --query "Effect vs RxJS" --strip-thinking

# Web search (sonar)
uv run .claude/skills/perplexity-tools/scripts/perplexity.py search --query "Nx 22 Crystal" --max-results 5

[1][OUTPUT]

Commands return: {"status": "success|error", ...}.

[INDEX][CMD][RESPONSE]
[1]ask{query, response, citations[]}
[2]research{query, response, citations[]}
[3]reason{query, response}
[4]search{query, results[]}