H
Howardism
Howardism · Vol. 03Plate I · No. 01

Security, tagged.

Notes42TagSecurityOldest28 May 2026Newest24 Sept 2026

Every article tagged security, newest first.

  1. C01
    AI Control vs. Alignment
    SecurityGovernanceIncident Response+2
    · 16′
  2. E02
    Google Threat Intelligence Group (GTIG)
    EntityOrgSecurity+1
    · 5′
  3. C03
    Harness Configuration Defects
    SecuritySupply ChainMCP+4
    · 17′
  4. E04
    Irregular
    EntityOrgAI Evaluation+2
    · 5′
  5. C05
    The Stolen Model-Access Economy
    SecurityThreat LandscapeSupply Chain+2
    · 17′
  6. C06
    Agent Self-Poisoning (the CREATE-Path)
    SecuritySupply ChainPropagation+3
    · 46′
  7. S07
    Memory-Poisoning Numbers, Conditioned on the Write
    DerivedSecurityMemory+1
    · 26′
  8. C08
    Mind Viruses (Agent-to-Agent Idea Propagation)
    SecurityMulti AgentPropagation+3
    · 40′
  9. C09
    Observability-Pipeline Poisoning
    SecurityPrompt InjectionObservability+2
    · 24′
  10. C10
    The OpenAI / Hugging Face Intrusion (July 2026)
    SecurityIncident ResponseThreat Landscape+2
    · 68′
  11. C11
    Remote MCP Authentication in the Wild
    SecurityMCPOauth+3
    · 29′
  12. C12
    Unsanctioned Agent Message Boards
    SecurityMulti AgentCovert Channel+3
    · 50′
  13. S13
    Guarantees That Degrade at Deployment: Action-Space Soundness, Admissibility Without Effect, and a Vendor-Coupled Security Framework
    DerivedAgent EngineeringSecurity+3
    · 34′
  14. C14
    Unsanctioned Action in Capability Evaluations
    SafetyAlignmentEvaluation+3
    · 73′
  15. C15
    Self-Propagating Prompt Injection (AI Worms)
    SecurityPrompt InjectionThreats+2
    · 31′
  16. C16
    Write-Then-Trusted
    SecuritySandboxingTrust Boundary+2
    · 44′
  17. C17
    Autonomous Intrusion
    SecurityIncident ResponseThreat Landscape+3
    · 52′
  18. S18
    Bind, Don't Forbid; Prevent, Don't Detect: The Action-Open and Poisoned-Memory Residuals
    DerivedSecurityPrompt Injection+2
    · 10′
  19. S19
    Classifier Gates vs OS Sandboxing: The Defense-in-Depth Story for Auto Mode and Cowork
    DerivedSecurityClaude Code+2
    · 14′
  20. S20
    Does 'Impossible, Not Tedious' Kill Defense-in-Depth? Layered Friction, Agent-Relativity, and the Frequency Paradox
    DerivedSecurityZero Trust+2
    · 13′
  21. C21
    Security Debt of Agent-Generated Code
    SecurityCode QualityCode Review+2
    · 29′
  22. C22
    Capability Gating Is Not Authorization
    SecurityAuthorizationConfused Deputy+3
    · 71′
  23. S23
    Can Models Learn to Separate Instructions from Data? Durable Property vs Training Gap
    DerivedSecurityPrompt Injection+2
    · 11′
  24. C24
    MCP Tool Poisoning
    SecurityMCPTool Poisoning+2
    · 45′
  25. C25
    Non-Malleable Memory Authority (TMA-NM)
    SecurityMemoryInformation Flow Control+2
    · 49′
  26. C26
    Off-Host, Identity-Bound Authorization
    SecurityAuthorizationIdentity+3
    · 31′
  27. C27
    Task-Specification Effects in Prompt Injection (AutoDojo)
    SecurityPrompt InjectionAdaptive Evaluation+2
    · 19′
  28. C28
    Agent Data Injection (ADI)
    SecurityPrompt InjectionThreats+2
    · 31′
  29. C29
    Agent Identity Management System (AIMS)
    SecurityIdentityAuthentication+3
    · 38′
  30. C30
    Out-of-Band Prompt-Injection Defense
    SecurityPrompt InjectionReference Monitor+2
    · 75′
  31. S31
    Foundation → Enterprise → Advanced: Is the Agent Access-Control Jump a Cliff?
    DerivedSecurityZero Trust+2
    · 8′
  32. C32
    Agent Identity and Authentication
    SecurityIdentityAuthentication+1
    · 31′
  33. C33
    Agent Supply Chain Risk
    SecuritySupply ChainMCP+2
    · 57′
  34. C34
    Agentic Prompt Injection
    SecurityPrompt InjectionThreats+1
    · 45′
  35. C35
    AI-Accelerated Offense
    SecurityThreat LandscapeVulnerability Research+1
    · 32′
  36. C36
    Autonomous Defense
    SecuritySoarIncident Response+2
    · 38′
  37. C37
    Blast Radius (Agentic)
    SecurityIsolationContainment+1
    · 52′
  38. C38
    Impossible, Not Tedious (Design Test)
    SecurityDesign PrincipleZero Trust+1
    · 24′
  39. C39
    Least Agency
    SecurityLeast PrivilegeAccess Control+1
    · 18′
  40. C40
    Memory and Context Poisoning
    SecurityMemoryRAG+1
    · 88′
  41. E41
    OWASP
    EntityOrgSecurity+1
    · 3′
  42. C42
    Zero Trust for AI Agents
    SecurityZero TrustAgent Deployment+1
    · 25′