相关职业
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/duc01226/EasyPlatform --skill deep-research命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
正在显示 SKILL.md
[Architecture] Use when designing solution architecture across backend, frontend, deployment, monitoring, testing, and code quality.
[Utilities] Use when you need to answer technical and architectural questions.
[Content] Use when you need to brainstorm as a PO/BA — structured ideation for problem-solving, new product creation, or feature enhancement.
基于 SOC 职业分类
| name | deep-research |
| version | 1.0.0 |
| description | [Research] Use when deeply researching top sources from web-research. |
[BLOCKING] Execute skill steps in declared order. NEVER skip, reorder, or merge steps without explicit user approval. [BLOCKING] Before each step or sub-skill call, update task tracking: set
in_progresswhen step starts, setcompletedwhen step ends. [BLOCKING] Every completed/skipped step MUST include brief evidence or explicit skip reason. [BLOCKING] If Task tools are unavailable, create and maintain an equivalent step-by-step plan tracker with the same status transitions.
Goal: Deep-dive into top sources to produce a cross-validated, source-cited evidence base (_evidence-{slug}.md) where every finding carries a confidence score, traces to specific sources, and flags discrepancies — never an unverified single-source claim presented as fact.
Summary:
.claude/tmp/_sources-{slug}.md) and turns prioritized Tier 1-2 sources into structured findings — it is not a fresh search..claude/tmp/_evidence-{slug}.mdWorkflow:
Key Rules:
Be skeptical. Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence percentages (Idea should be more than 80%).
Web Research Protocol — Every factual claim needs 2+ independent sources. Source tiers: Tier 1 (authoritative .gov/.edu/official docs), Tier 2 (industry reports), Tier 3 (credible blogs — cross-validate), Tier 4 (unverified — NEVER cite as fact). Declare confidence (95/80/60/<60%) for all findings. Working files →
.claude/tmp/, final output →docs/knowledge/. Canonical protocol lives in theweb-researchskill.
Read the source map from .claude/tmp/_sources-{slug}.md (output of web-research step).
Prioritize sources for deep-dive:
For each priority source (max 8):
WebFetch with the URLFor each source, extract:
Compare findings across sources:
Write to .claude/tmp/_evidence-{slug}.md:
# Evidence Base: {Topic}
**Date:** {date}
**Sources analyzed:** {count}
## Findings
### Finding 1: {Title}
**Confidence:** {95%|80%|60%|<60%}
**Sources:** [1], [3]
**Content:** {finding with inline citations}
**Cross-validation:** {agreement/discrepancy notes}
## Unresolved Discrepancies
- {claim X from source A vs claim Y from source B}
## Gaps Remaining
- {what couldn't be verified}
MANDATORY IMPORTANT MUST ATTENTION — NO EXCEPTIONS: If you are NOT already in a workflow, you MUST ATTENTION use
AskUserQuestionto ask the user. Do NOT judge task complexity or decide this is "simple enough to skip" — the user decides whether to use a workflow, not you:
- Activate
workflow-researchworkflow (Recommended) — web-research → deep-research → synthesis → review- Execute
/deep-researchdirectly — run this skill standalone
MANDATORY IMPORTANT MUST ATTENTION — NO EXCEPTIONS after completing this skill, you MUST ATTENTION use AskUserQuestion to present these options. Do NOT skip because the task seems "simple" or "obvious" — the user decides:
[IMPORTANT] Use
TaskCreateto break ALL work into small tasks BEFORE starting.
External Memory: For complex or lengthy work (research, analysis, scan, review), write intermediate findings and final results to a report file in
plans/reports/— prevents context loss and serves as deliverable.
Evidence Gate: MANDATORY IMPORTANT MUST ATTENTION — every claim, finding, and recommendation requires
file:lineproof or traced evidence with confidence percentage (>80% to act, <80% must verify first).
AI Mistake Prevention — Failure modes to avoid on every task:
Re-read files after context changes. Context compaction, resume, or long-running work can make memory stale; verify current files before acting. Verify generated content against source evidence. AI hallucinates APIs, names, claims, and document facts. Check the relevant source before documenting or referencing. Check downstream references before deleting or renaming. Removing an artifact can stale docs, generated mirrors, configs, and callers; map references first. Trace the full impact chain after edits. Changing a definition can miss derived outputs and consumers. Follow the affected chain before declaring done. Verify ALL affected outputs, not just the first. One green check is not all green checks; validate every output surface the change can affect. Assume existing values are intentional — ask WHY before changing. Before changing a constant, limit, flag, wording, or pattern, read nearby context and history. Surface ambiguity before acting — don't pick silently. Multiple valid interpretations require an explicit question or stated assumption with risk. Keep shared guidance role-relevant. Universal guidance must help every receiving skill or agent; code-specific obligations belong only in code-specific protocols.
Critical Thinking Mindset — Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence >80% to act. Anti-hallucination: Never present guess as fact — cite sources for every claim, admit uncertainty freely, self-check output for errors, cross-reference independently, stay skeptical of own confidence — certainty without evidence root of all hallucination.
MUST ATTENTION apply critical + sequential thinking — every claim needs appropriate traced evidence (file:line for repo/code claims; source URL or artifact section for research, product, content, and docs claims); confidence >80% to act, <60% DO NOT recommend. Anti-hallucination: never present guess as fact, admit uncertainty freely, cross-reference independently, stay skeptical of own confidence.
MUST ATTENTION apply AI mistake prevention — verify generated content against evidence, trace downstream references before deleting or renaming, verify all affected outputs, re-read files after context loss, and surface ambiguity before acting.
IMPORTANT MUST ATTENTION follow declared step order for this skill; NEVER skip, reorder, or merge steps without explicit user approval
IMPORTANT MUST ATTENTION for every step/sub-skill call: set in_progress before execution, set completed after execution
IMPORTANT MUST ATTENTION every skipped step MUST include explicit reason; every completed step MUST include concise evidence
IMPORTANT MUST ATTENTION if Task tools unavailable, maintain an equivalent step-by-step plan tracker with synchronized statuses
IMPORTANT MUST ATTENTION Goal: Produce a cross-validated, source-cited evidence base (_evidence-{slug}.md) where every finding carries a confidence score, traces to specific sources, and flags discrepancies — never an unverified single-source claim presented as fact.
IMPORTANT MUST ATTENTION — Protocols in force (concise digest of the SYNC/shared blocks this skill carries):
IMPORTANT MUST ATTENTION every finding cites a specific source by number; conflicting claims → present BOTH positions, flag as discrepancy; lone source → mark "single source, unverified" — why: an uncited or uncross-checked claim presented as fact is the failure this skill exists to prevent.
IMPORTANT MUST ATTENTION cross-validation drives confidence — 2+ independent sources agree = high (95/80%), 1 source = "unverified", disagreement = discrepancy with both sides; NEVER collapse a conflict into one tidy answer — why: hidden conflicts ship as false certainty downstream.
IMPORTANT MUST ATTENTION declare a confidence percentage (95/80/60/<60%) on EVERY finding; <60% evidence DO NOT present as fact — say "insufficient evidence, verified: … / not verified: …" instead.
IMPORTANT MUST ATTENTION cap WebFetch at 8 calls per invocation; spend them on Tier 1-2 authoritative sources covering identified gaps, NEVER Tier 4 unverified content as fact — why: budget discipline forces prioritization over breadth.
IMPORTANT MUST ATTENTION this is the deep-DIVE stage — consume the prior _sources-{slug}.md map; do NOT start a fresh search — why: the source map already triaged and tiered candidates, re-searching wastes the WebFetch budget.
IMPORTANT MUST ATTENTION capture per source: publication date, author credentials, source type, methodology — why: confidence must be defendable later, not asserted from memory.
IMPORTANT MUST ATTENTION the deliverable MUST include an ## Unresolved Discrepancies section and a ## Gaps Remaining section — NEVER hide what couldn't be verified.
IMPORTANT MUST ATTENTION verify AI-generated facts/quotes/numbers against the actual fetched source before recording — NEVER fabricate a citation, stat, or quote — why: a hallucinated source corrupts the whole evidence base silently.
IMPORTANT MUST ATTENTION break work into small TaskCreate todos BEFORE starting; keep one in_progress; add a final review todo verifying every finding is cited and confidence-scored.
IMPORTANT MUST ATTENTION write intermediate findings incrementally to .claude/tmp/_evidence-{slug}.md (External Memory) — NEVER hold the full evidence base in context only — why: context loss before the final write loses all extracted findings.
IMPORTANT MUST ATTENTION validate route decisions with the user via AskUserQuestion — never auto-decide whether to run the workflow vs. this skill standalone.
Anti-Rationalization:
| Evasion | Rebuttal |
|---|---|
| "One good source is enough" | A lone source is "single source, unverified" — never high confidence. Cross-validate. |
| "The sources roughly agree, call it settled" | Roughly ≠ exactly. Record the discrepancy with both positions; don't smooth it over. |
| "I remember this stat from the page" | Re-open the fetched source and verify the number/quote before citing. Memory hallucinates. |
| "I'll fetch a few more to be thorough" | 8-call cap is the budget. Prioritize Tier 1-2 gap-coverage, not breadth. |
| "I'll write the evidence base at the end" | Persist findings incrementally to _evidence-{slug}.md — a context cutoff loses batched work. |
IMPORTANT MUST ATTENTION every finding cites a specific source + carries a confidence % (95/80/60/<60%); conflicts → both positions flagged, lone source → "unverified".
IMPORTANT MUST ATTENTION cap WebFetch at 8 Tier 1-2 calls and persist the evidence base incrementally to .claude/tmp/_evidence-{slug}.md.
IMPORTANT MUST ATTENTION the deliverable must surface ## Unresolved Discrepancies and ## Gaps Remaining — never hide what couldn't be verified.
[IMPORTANT] Use
TaskCreateto break ALL work into small tasks BEFORE starting; add a final review todo to verify work quality.