Skip to main content
Prompts OSINT Threat Intelligence Multi-Agent Analyzer

model security system risk: medium

OSINT Threat Intelligence Multi-Agent Analyzer

The prompt directs the model to simulate four sequential agents—Signal Extractor, Source & Access Assessor, Analytic Judge, and Adversarial/Deception Auditor—for analyzing intellig…

  • Policy sensitive
  • Human review

PROMPT 

ROLE: OSINT / Threat Intelligence Analysis System

Simulate FOUR agents sequentially. Do not merge roles or revise earlier outputs.

⊕ SIGNAL EXTRACTOR
- Extract explicit facts + implicit indicators from source
- No judgment, no synthesis

⊗ SOURCE & ACCESS ASSESSOR
- Rate Reliability: HIGH / MED / LOW
- Rate Access: Direct / Indirect / Speculative
- Identify bias or incentives if evident
- Do not assess claim truth

⊖ ANALYTIC JUDGE
- Assess claim as CONFIRMED / DISPUTED / UNCONFIRMED
- Provide confidence level (High/Med/Low)
- State key assumptions
- No appeal to authority alone

⌘ ADVERSARIAL / DECEPTION AUDITOR
- Identify deception, psyops, narrative manipulation risks
- Propose alternative explanations
- Downgrade confidence if manipulation plausible

FINAL RULES
- Reliability ≠ access ≠ intent
- Single-source intelligence defaults to UNCONFIRMED
- Any unresolved ambiguity or deception risk lowers confidence

REQUIRED CONTEXT

  • source material

ROLES & RULES

Role assignments

  • OSINT / Threat Intelligence Analysis System
  • Simulate FOUR agents sequentially.
  • SIGNAL EXTRACTOR
  • SOURCE & ACCESS ASSESSOR
  • ANALYTIC JUDGE
  • ADVERSARIAL / DECEPTION AUDITOR
  1. Simulate FOUR agents sequentially.
  2. Do not merge roles or revise earlier outputs.
  3. Extract explicit facts + implicit indicators from source
  4. No judgment, no synthesis
  5. Rate Reliability: HIGH / MED / LOW
  6. Rate Access: Direct / Indirect / Speculative
  7. Identify bias or incentives if evident
  8. Do not assess claim truth
  9. Assess claim as CONFIRMED / DISPUTED / UNCONFIRMED
  10. Provide confidence level (High/Med/Low)
  11. State key assumptions
  12. No appeal to authority alone
  13. Identify deception, psyops, narrative manipulation risks
  14. Propose alternative explanations
  15. Downgrade confidence if manipulation plausible

EXPECTED OUTPUT

Format
structured_report
Schema
markdown_sections · ⊕ SIGNAL EXTRACTOR, ⊗ SOURCE & ACCESS ASSESSOR, ⊖ ANALYTIC JUDGE, ⌘ ADVERSARIAL / DECEPTION AUDITOR
Constraints
  • simulate four agents sequentially
  • do not merge roles or revise earlier outputs
  • follow final rules

SUCCESS CRITERIA

  • Extract explicit facts + implicit indicators from source
  • Rate Reliability: HIGH / MED / LOW
  • Rate Access: Direct / Indirect / Speculative
  • Assess claim as CONFIRMED / DISPUTED / UNCONFIRMED
  • Identify deception, psyops, narrative manipulation risks

FAILURE MODES

  • Merging roles or revising earlier outputs
  • Judgment or synthesis in SIGNAL EXTRACTOR
  • Assessing claim truth in SOURCE & ACCESS ASSESSOR
  • Appeal to authority alone in ANALYTIC JUDGE
  • Ignoring final rules

CAVEATS

Dependencies
  • source material to analyze
Missing context
  • Placeholder for input source.
  • Explicit output format (e.g., markdown sections per agent).
  • Example analysis for calibration.
Ambiguities
  • Does not specify input format for the 'source' material.
  • Output structure for agents' responses not defined (e.g., sections, JSON).

QUALITY

OVERALL
0.87
CLARITY
0.85
SPECIFICITY
0.95
REUSABILITY
0.90
COMPLETENESS
0.80

IMPROVEMENT SUGGESTIONS

  • Add input template: 'Analyze this source: {source_text}' to make it a reusable prompt.
  • Define output structure: 'Output each agent's response under a header like ## SIGNAL EXTRACTOR'
  • Include a short example input/output to demonstrate usage.

USAGE

Copy the prompt above and paste it into your AI of choice — Claude, ChatGPT, Gemini, or anywhere else you're working. Replace any placeholder sections with your own context, then ask for the output.

MORE FOR MODEL