菜单
Open-Source Intelligence research methodology -- multi-wave, progressively deepening investigation with structured intel output
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。
基于 SOC 职业分类
| name | osint |
| description | Open-Source Intelligence research methodology -- multi-wave, progressively deepening investigation with structured intel output |
| version | 1.0.0 |
| triggers | ["osint on","do osint","investigate","intelligence on","research on","dig into","deep dive on"] |
Purpose: Conduct structured, multi-wave open-source intelligence research on a subject (person, organization, event, or topic) using publicly available sources. Produce a formatted intelligence briefing with sourced analysis and prediction matrices.
Pattern: The Investigative Spiral -- each wave narrows focus, increases depth, and builds on prior findings.
This skill is for research using publicly available information only:
Never:
Goal: Establish baseline profile. Cast wide net.
Launch 3-4 parallel WebSearch queries covering different angles:
WebSearch: "$SUBJECT profile background biography"
WebSearch: "$SUBJECT latest news [current year]"
WebSearch: "$SUBJECT [domain] structure leadership"
WebSearch: "$SUBJECT [government/legal] status charges"
Output: Basic identity, timeline, key relationships, current status.
Goal: Drill into key figures, relationships, and dynamics identified in Wave 1.
Launch 3-4 parallel WebSearch queries on specific findings:
WebSearch: "[Key Figure A] role connection to $SUBJECT"
WebSearch: "[Organization] structure members sanctions"
WebSearch: "[Event/Conflict] details timeline analysis"
WebSearch: "[Financial/Legal] aspect indictments court records"
Supplement with WebFetch on high-value URLs found in Wave 1 (government profiles, think tank reports, court documents).
Output: Detailed profiles of key figures, org charts, relationship maps.
Goal: Get expert analysis, academic research, and predictive frameworks.
Target authoritative analytical sources:
WebSearch: "$SUBJECT [analytical angle] analysis expert assessment"
WebFetch: [think tank URL] "Extract all analytical details about..."
WebFetch: [government URL] "Extract complete profile/designation..."
Output: Expert predictions, historical parallels, structural analysis.
Goal: Go deep on whatever angle the user requests.
This wave is triggered by user direction ("dig deeper into X"). Apply the same multi-parallel search pattern but focused entirely on the requested angle.
Output: Exhaustive detail on the specific sub-topic.
Prioritize sources by reliability:
| Tier | Source Type | Examples | Trust Level |
|---|---|---|---|
| 1 | Government primary | DEA, State Dept, OFAC, court records | Highest (official) |
| 2 | Academic/Think tank | Brookings, ACLED, university research | High (peer-reviewed) |
| 3 | Quality investigative | InSight Crime, Bellingcat, OCCRP | High (methodology-driven) |
| 4 | Major news wire | AP, Reuters, AFP, NPR, PBS | Medium-high (editorial standards) |
| 5 | Regional/specialized | Domain-specific news outlets | Medium (domain expertise, less editorial rigor) |
| 6 | Social media/unverified | X/Twitter, Telegram, anonymous sources | Low (corroborate before using) |
Rule: Never state something as fact from Tier 6 alone. Always triangulate across tiers.
# OSINT [PROFILE/BRIEFING/ANALYSIS]: $SUBJECT
## 1. IDENTITY / OVERVIEW
[Basic profile, aliases, DOB, origin, physical description if known]
## 2. TIMELINE / HISTORY
[Chronological key events, rise to prominence, arrests, etc.]
## 3. ORGANIZATIONAL STRUCTURE
[Org chart table: Name | Alias | Role | Status | Reward]
[Key relationships, family connections, alliances]
## 4. OPERATIONAL ANALYSIS
[How they operate, territorial control, revenue streams, methods]
## 5. CURRENT DYNAMICS
[Breaking news, recent developments, ongoing conflicts]
## 6. SCENARIO ANALYSIS
[Prediction matrix with probability, timeline, triggers]
| Scenario | Probability | Timeline | Trigger |
## 7. KEY METRICS
[Numbers table: members, revenue, territory, etc.]
## 8. SOURCES
[All URLs as markdown hyperlinks, organized by tier]
| Problem | Recovery |
|---|---|
| WebFetch returns 403/blocked | Fall back to WebSearch for same info from different sources |
| Source contradicts another | Note the contradiction, cite both, assess which is more reliable |
| Paywall blocks key article | Search for the same story on wire services (AP/Reuters) or quote the preview |
| Limited English sources | Search in the subject's native language, translate key findings |
| Information is stale | Note the date, flag as potentially outdated, search for updates |
| Modifier | Values | Default | Description |
|---|---|---|---|
| @depth: | 1-4 | 3 | Number of research waves |
| @focus: | string | broad | Initial angle to emphasize |
| @output: | report, brief, profile | report | Output format (full report, executive brief, or identity profile) |
# Full OSINT research
/osint "subject name"
# Quick profile only
/osint @depth:1 @output:profile "target name"
# Focused angle
/osint @focus:finances "organization name"
Extracted from a real OSINT research session (Feb 2026). The session involved 4 progressive research waves, 27 parallel web searches, and 10 deep article extractions, producing 2 structured intelligence briefings with prediction matrices, org charts, and link analysis. The multi-wave parallel search methodology proved significantly more effective than sequential single queries.
Part of the Claude Code Skills collection -- skills extracted from real AI-assisted development sessions.
Install a lightweight inline-editing CMS into any Vercel-deployed project — content store + EditableText component + cookie-auth admin shell + Vercel KV/Upstash/Blob storage. Auto-detects the framework and installs natively for Next.js (App Router & Pages Router) and Vite + React. For SvelteKit, Astro, Nuxt, or plain static, prints a manual-port guide. Use when the user asks to add a CMS, inline editing, editable text, an admin panel, or "make this site editable."
Audit and improve a Claude Code memory layer end-to-end. Measures CLAUDE.md and MEMORY.md against Anthropic's documented caps (200 lines per CLAUDE.md, 200 lines OR 25 KB for MEMORY.md auto-load), surveys project portfolio on disk vs memory coverage, finds sync conflicts in ~/.claude/, identifies gaps tier-by-tier, and executes improvements under bias-to-action with cap-fitting verification at each step. Use when CLAUDE.md/MEMORY.md "feel bloated", when memory feels invisible in fresh sessions, or as periodic hygiene (~monthly).
Burn project bloat to reveal essence. Surgical, intent-anchored, multi-agent. Articulate a three-layer intent (business / architectural / stylistic), dispatch 4 parallel subagents to find code-rot / structural / dep / doc-test bloat, produce a tiered CALCINATION-PLAN.md, then execute per-item with auto-revert on failed verify. Reversible via git. Use when a project has gotten fat and needs aligning back to its core purpose.
A discipline for working with AI code generation. Replaces algorithm-writing with the meta-skill that Google's new Code Comprehension Interview tests: directing AI codegen, validating its output, and debugging what AI produces. Five principles + five triggers, with the discipline applied to Claude's own output as well as the user's. Use when actively pairing with AI on real code.
Opinionated meta-prompting command suite — 9 composable patterns built on categorical foundations. Use when the user wants to set up a meta-prompting workflow, route prompts by context, frame goals, build subagents, distill session learnings into durable artifacts, or invoke the alchemical 4-stage / Hermetic-correspondence patterns for personal/creative work.
Build native Android APKs without Android Studio using raw SDK tools (javac, d8, aapt, apksigner). Produces tiny APKs (~30KB) that build in under 2 seconds.