Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add ccoalm/ccl-skills --skill multi-perspective-researchgit clone --depth 1 https://github.com/ccoalm/ccl-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/ccoalm/ccl-skills/multi-perspective-research)<a href="https://agentmods.dev/skills/ccoalm/ccl-skills/multi-perspective-research"><img src="https://agentmods.dev/badge/skills/ccoalm/ccl-skills/multi-perspective-research/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/ccoalm/ccl-skills/multi-perspective-research"><img src="https://agentmods.dev/badge/skills/ccoalm/ccl-skills/multi-perspective-research.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00228 | $0.09379 |
| Opus 5 | $0.00114 | $0.04689 |
| Sonnet 5 | $0.00046 | $0.01876 |
| Haiku 4.5 | $0.00023 | $0.00938 |
Grade A, and why
multi-perspective-research scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 71 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Multi-Perspective Research（多视角主题调研）
当用户要研究、调研、深入了解一个主题（写文档前、竞品/组织/领域研究、演讲/谈判准备，或被 product-rd-workflow 的选型/评估 gate 调用产调研底稿）时使用。单问单答只会返回最大众化的表面共识；本技能用结构化的多视角流程逼出争议、利益结构、历史先例和领域盲区。
方法改编自 STORM（Shao et al., "Assisting in Writing Wikipedia-like Articles From Scratch with Large Language Models", NAACL 2024；开源实现 github.com/stanford-oval/storm）。论文报告的同行评审评测显示，其成文条理性与覆盖面均显著优于 outline-driven RAG 基线（具体数字以原文为准：aclanthology.org/2024.naacl-long.347）。后续 Co-STORM（Jiang et al., EMNLP 2024）加入 moderator（从检索到但未被使用的信息中提问，专挖盲区）与动态思维导图。STORM 的核心洞察：研究质量取决于提问的视角多样性，而不是单个问题的措辞。
主张边界（外部证据核查后收窄）：证据支持多视角提问提升调研的覆盖面、条理性和盲区发现（写作/研究合成场景）；证据不支持"同一模型模拟多个 persona 辩论就能稳定提升事实正确性"——两项独立评测（Smit et al., "Should we be going MAD?", ICML 2024, PMLR v235；及 "Stop Overvaluing Multi-Agent Debate", arXiv:2502.08788, 2025）均显示同质 persona 辩论常打不过 self-consistency 等单模型基线，模型异构与真实检索才是可靠增益。所以本技能把多视角用于提问与覆盖，把正确性交给下面的证据落地纪律。
何时不用
- 用户已有方案、要的是追问和压力测试 →
grill-me。 - 请求本质是交付决策（选型结论、可行性评估、方案评估）→
product-rd-workflow的对应 gate 拥有决策与证据标准；本技能只产出调研底稿。判别标准取可操作面：请求点名候选让选或用裁决措辞要结论（选哪个/要不要/该不该/可不可行）→product-rd-workflow；纯主题/领域/写作准备调研 → 本技能。调研中若演变为要拍板（用户开始追问"那选哪个"），当场移交product-rd-workflow。 - 一次性的简单事实问题（一查便知）→ 直接回答，不要套流程。
核心流程
先做覆盖预检，再按顺序执行四步。轻量场景（快速了解）可只跑 1+3，但此模式不得输出"可操作建议"或任何载重结论，且未跑矛盾图时发现列表不得声称"按可靠性排序"（可靠性判断依赖矛盾图/自评）——产出只能作为探索性概览交付，仍必须附逐条 grounded/ungrounded 标注；任何包含可操作建议/载重结论/可靠性排序的简报必须跑完第 2、4 步并附待核验清单。提示词模板见 references/research-prompts.md。
- 覆盖预检：第一次检索前先写覆盖矩阵：
研究问题/载重主张 | 必需证据类别 | 检索入口 | 关闭条件 | 状态。证据类别由主张决定，至少区分用户提供的线索、一手/官方材料、独立外部证据和反例/负面证据；主张依赖时效、规模或因果时，再增加对应类别。仅当用户确实未提供任何材料或线索时，该类才能记不适用，并必须写明原因及其对结论的限制；用户已提供任何材料时不得标不适用，也不得为凑齐四类虚构线索。这四类是完整/深度标签的组织控制，不是宣称所有主题都遵循同一套系统综述标准。用户材料默认只是线索，不是研究边界；只有用户明确要求“仅基于这些材料”时才能封闭外部检索。“本轮不要执行检索/先给计划”只把检索状态留为pending，不等于限定来源范围。每格必须以本次取回的来源关闭，或写明不可得/不适用及其对结论的限制。矩阵未闭合时，只能交付定向/部分调研，不得宣称“深度调研完成”。外部方法依据与适用边界见../skill-extraction-workflow/references/external-practice-controls.md#research-coverage。调研对象是一个具体组织（而非主题）时，先定 estimand 再走一遍组织披露渠道类别清单：四类证据问的是"证据是什么种类"，只查官网与年报就能把"一手/官方"格正当关上，而强制披露制度、监管方批量数据、他国法定登记通常信息量更大却一次未被想到。references/public-disclosure-channels.md是这个组织场景的渠道起点，只用来防"没想到"，是起点不是全集，不另立闭合状态；走完它不构成任何终局判断，走不动的类别按本节既有纪律计为未关闭格。references/public-data-acquisition.md则说明公开材料怎么取回、解析与自校验，适用于组织、主题、政策、技术、市场等场景。组织场景的 estimand 要写清目标是单一法人、合并集团、品牌或业务单元还是某个网络、依据与截至日期；依据不足时只能称"有界的工作口径"，且人数与比例类结论不得进入承重结论。
What ships with it
6 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 9d ago First seen · 71 lines · 228 tokens per session scan A b10d8b03bea9
multi-perspective-research is a skill published in the GitHub repository ccoalm/ccl-skills (6 stars, last pushed today), licensed Apache-2.0. It adds 228 tokens to every session and 9,379 once invoked, about $0.0011 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
security-review
Perform a focused security review of pending git changes to identify high-confidence security vulnerabilities with real exploitation potential. Use this skill when the user asks for a security review, security audit, vulnerability scan, or wants to check pending changes on a branch for security issues before merging.…
huggingface-llm-trainer
Train or fine-tune language models with TRL or Unsloth on Hugging Face Jobs, including SFT, DPO, GRPO, reward models, and GGUF conversion. Use for cloud LLM training; use huggingface-vision-trainer for vision tasks.
feature-dev
Guide a feature implementation through a structured seven-phase workflow with deep codebase understanding, clarifying questions, parallel architecture design, and quality review. Use this skill when the user asks to build a new feature, add functionality, or wants a methodical approach to implementation rather than…
memstack-business-gdpr
Use this skill when the user says 'GDPR', 'data protection', 'privacy compliance', 'DPA', 'DSAR', 'data subject request', 'cookie consent', 'privacy audit', 'CCPA', or asks 'do I need GDPR for this repo'. Scans the repository to detect what personal data is collected, classifies sensitivity, determines whether GDPR…
memstack-deployment-domain-ssl
Use this skill when the user says 'setup domain', 'configure DNS', 'SSL certificate', 'domain-ssl', 'custom domain', 'HTTPS setup', or needs to configure DNS records, SSL certificates, and custom domains for any hosting provider. Do NOT use for full deployment workflows.
memstack-development-refactor-planner
Use this skill when the user says 'refactor', 'refactoring plan', 'code cleanup', 'reduce duplication', 'simplify code', 'tech debt', 'god class', 'tight coupling', or needs to systematically improve existing code. Identifies targets, assesses risk, and builds incremental execution plans. Do NOT use for writing new…