[un]prompted March 2026 Talks — Sorted by Relevance to RA + CMM

Naming

The source page is headlined “[un]prompted II” but the body says “back for the second time in (or around) September” — i.e. the September 2026 event is [un]prompted II, and this March 3–4, 2026 conference is the prior one. This wiki uses date-based slugs to avoid the ambiguity. See [[unprompted-conference-march-2026|[un]prompted Conference — AI Security Practitioner Conference (March 3–4, 2026)]] for the catalog.

What this page is for. Anton wants to track down slides/notes/videos for the talks most likely to feed the six-plane RA and the 9-domain CMM. This page ranks the entire agenda from highest to lowest relevance to Securing AI (security OF agentic AI in enterprise), with the reasons made explicit so the ranking is auditable.

Scope filter. “Securing AI” = controls and architecture for protecting AI agents and AI-using systems. Talks about “AI for security” (using AI to find traditional vulns, do DFIR, build SOC tooling) are separated below, not omitted, because some of them still inform CMM domains (D7 Observability & Detection, D9 Operations).

Tier 1 — Direct evidence for one or more RA planes / CMM domains

These talks present concrete production architecture, control patterns, or measurement methodology that should map directly into the RA, CMM evidence rubric, or the standards crosswalk. Highest priority for slide/video pursuit.

#TalkSpeaker(s)RA plane(s)CMM domain(s)Why it’s Tier 1
1Breaking the Lethal Trifecta (Without Ruining Your Agents) ✅ ingestedAndrew Bullen, StripeEgress, Control, ObservabilityD3, D4, D5Production Lethal Trifecta containment: Smokescreen + agent-tag CI; Toolshed + ToolAnnotations; queued/batched/optimistic confirmations; coins Lethal Bifecta. L3/L4 evidence; deep-agent enforcement still WIP per speaker.
2Guardrails beyond Vibes: Shipping Security Agents in ProductionJ. Zhang + S. Shah, StripeControl, Runtime, ObservabilityD3, D4, D7Orchestrator/child agent architecture + offline+online eval — directly slot into Oversight Layer (PDP + PEP for Agentic AI) and L3 evidence
3Capability-Based Authorization for AI Agents — Warrants That Survive Prompt Injection ✅ ingestedNiki Aimable Niyikiza, Tenuo / SnapIdentity, ControlD2, D3Macaroon/UCAN-lineage cryptographic warrants with monotonic attenuation across delegation hops; 4 deployment modes (in-process / sidecar / gateway / MCP-proxy); reports baseline 90%→0% multi-agent ASR on custom harness. Partially closes PEP gap from the delegation-aware angle.
4Hooking Coding Agents with Cedar ✅ ingestedMatt Maisel, SonderaControl, RuntimeD3, D4Open-source Cedar + Rust hook reference monitor for Cursor/Claude Code/Gemini CLI — trajectory event model (action/observation/control/state) as structural unit; IFC taint tracking across turns; policy agent for formal Cedar generation/validation over MCP; partially closes PEP gap from the hook-based angle
5Securing Workspace GenAI at Google Speed: Surviving the Perfect Storm ✅ ingestedNicolas Lidzborski, GoogleRuntime, Data, Egress, ControlD3, D4, D5, D7Three-year Google Workspace retrospective. Names Prompt as Code as the structural framing; Agency Gap / Orchestration Hijacking / Recursive Prompt Injection (and Semantic Gaslighting) as named threat sub-classes; the four-layer “Architecting the Fortress” structural blueprint; Plan-Validate-Execute Pattern as canonical HITL pattern; Sentinel Tokens (Prompt Delimitation) for prompt delimitation. Calendar-invite hijack (Nassi et al.) as worked example. Architectural sibling to Bullen’s Stripe talk (input/orchestration/output side; Bullen covers egress/tool-policy side).
6Your Agent Works for Me NowJohann RehbergerRuntime, Egress, ObservabilityD4, D5, D7Promptware as a class — delayed tool invocation, intent-activation evasion, persistence + lateral movement + C2 + exfil; previously undisclosed Gemini/Copilot exploits — feeds OWASP-ASI mappings in the CMM crosswalk
7Glass-Box Security: Operationalizing Mechanistic InterpretabilityCarl Hurd, StarseerObservabilityD7Internal EDR” + neuron-level semantic tripwires — adds a new D7 Glass-Box L4/L5 evidence path beyond OTel/SIEM
8Are Your LLM’s Safety Mechanisms Intact? Detecting Backdoors with White-Box AnalysisAkash Mukherje, Realm LabsData (model), ObservabilityD6, D7, D8Backdoors that pass black-box eval but fail white-box; relevant to Supply Chain Security for Agentic AI for model-supply-chain integrity
9Total Recon: How We Discovered 1000s of Open Agents in the WildA. Efrat + R. Ben Chaim, ZenityIdentity, Egress, ObservabilityD7, D8External-attack-surface evidence at scale; PowerPwn as an AI-BOM adjunct (recon tool to discover your own exposure)
10Building Secure Agentic Systems: Lessons from Daily-Driver AgentsBrooks McMillin, DropboxIdentity, Control, DataD2, D3, D6Multi-agent memory-isolation failures + the fix is a CMM evidence gem; capability bounding as concrete L3
111.8M Prompts, 30 Alerts: Hunting Abuse in a User-Defined Agent EcosystemM. Rittinghouse + M. Huang, SalesforceObservabilityD7Asset Rarity + Query Complexity behavioral baselines at production scale — concrete L4 evidence for D7 Agent Observability (insider-threat framing)
12Beyond the Chatbot: Delivering an Agentic SOCP. Smith + R. Sharma, SalesforceControl, ObservabilityD7, D9Polyphonic (supervisor–worker) architecture — direct architectural alternative; complements the RA’s Oversight Layer (PDP + PEP for Agentic AI)
13Training BrowseSafeKyle Polley, PerplexityRuntimeD4Production prompt-injection detection; F1 ~0.91 at <100ms; BrowseSafe-Bench is a CMM-friendly artifact (measurement + flywheel)
14The Parseltongue Protocol — 100+ Obfuscation MethodsJoey Melo, CrowdStrikeRuntime, DataD4, D6Largest published obfuscation eval (17,000+ prompts × 9 models) — drives D4 jailbreak-resistance L3/L4 evidence
15When Passports Execute: Exploiting AI-Driven KYC PipelinesSean Park, TrendAIData, RuntimeD4, D6Document-embedded Indirect Prompt Injection hitting compliance verification chains — concrete RAG Hardening failure mode
16Can You See What Your AI Saw? GenAI Endpoint ObservabilityMika Ayenson, ElasticObservabilityD7Argues for extending OpenTelemetry semantic conventions to GenAI tool activity — likely future standard input to D7 maturity rubric
17Detecting GenAI Threats at Scale with YARA-Like Semantic Rules (SYARA)Mohamed Nabeel, Palo Alto NetworksRuntime, ObservabilityD4, D798% detection at <100ms/$0.001 — economic argument that pure-LLM gates are wrong shape for D7
18Operation Pale Fire: Red-Teaming gooseW. Ring + J. Peedikayil, BlockRuntime, ObservabilityD4, D7, D9Self-red-team of an open-source agent — directly informs Red Teaming Capability Framework and CMM L3 internal-red-team evidence
19Rethinking How We Evaluate Security AgentsMudita Khurana, AirbnbObservabilityD7, D9Find→confirm→patch→validate loop + capability-centric eval — directly extends CLASP
20The AI Security Larsen Effect (capability-based vendor framework)Maxim Kovalsky, Consortium Networks(cross-cutting)D1, D9Capability-based configure/buy/build framework — usable in CMM L2 vendor-shortlist evidence
21AI Security with GuaranteesIlia Shumailov, AI SequrityControl, RuntimeD3, D4Claims security guarantees for computer-use agents; if delivered, opens an L4/L5 evidence path
22The Advent of Confidential AIRaghu Yeluri, IntelData, RuntimeD4, D6TEEs for inference + prompts + context — concrete control for D6 (model isolation) and the Insight map’s “model security” stripe
23Trajectory-Aware Post-Training of Open-Weight Models for Security AgentsA. Brown + M. Prashant, AWSData (model)D6, D9Open-source SFT→GRPO pipeline + reward design — concrete approach to “build the security agent yourself” rather than depending on frontier API
24Vibe Check: Security Failures in AI-Assisted IDEsPiotr Ryciak, MindgardRuntime, EgressD4, D5, D8Catalog of zero/one-click + autorun + delayed-trigger paths in Codex, Kiro, Antigravity, Cursor — direct Supply Chain Security for Agentic AI evidence
25Injecting Security Context During Vibe CodingSrajan Gupta, DaveControl, RuntimeD3, D4, D8MCP server that injects security guidance pre-generation + verifies post-generation — slots into MCP Security + L3 evidence

Tier 2 — Adjacent: production agent threats, governance, AI-assisted incidents

Useful for the RA threat-control matrix and for situating CMM thresholds; less directly an architectural artifact.

#TalkSpeaker(s)Why Tier 2
268 Minutes to Admin / EtherRAT (VibeHacking)Sergej Epp, SysdigTwo in-the-wild AI-assisted attack campaigns — feeds an incidents page on AI-assisted ops; behavioral attribution methodology for AI-assistance
27SIFT — FIND EVIL!! Claude Code on DFIR SIFTRob T. Lee, SANSOperational MCP wiring + reference to Anthropic GTG-1002 (80–90% autonomous adversaries) — important context, mostly defender-side AI tooling
28Three Phases of AI AdoptionChase Hasbrouck, US Army Cyber CommandAdoption-curve framing useful for D9 maturity narrative (Shadow Automation supports)
29Establishing AI Governance Without Stifling InnovationBilly Norwood, FFF EnterprisesHealthcare AI governance committee — D1 governance evidence
30Enterprise AI Governance at SnowflakeRagini Ramalingam, SnowflakeEnterprise governance program — D1 governance evidence
31200 Bugs/Week/Engineer: Trail of BitsDan Guido, Trail of Bits”AI-native operating system” of incentives + sandboxing — D9 organizational redesign + L4 evidence narrative; provocative claim worth scrutiny
32Kinetic Risk: Securing & Governing Physical AIPadma Apparao, IntelVLA model risks — extends scope beyond text agents; useful for “embodied” annex to RA/CMM
33AI Notetakers: The Most Important Person in the RoomJoe SullivanAI-as-meeting-record governance gap — D1, D9 (consent + discovery)
34Code Is Free: Securing Software in the Agentic FutureP. McMillan + R. Lopopolo, OpenAIEngineering-first invariants — useful framing, mostly principle-level
35The Hard Part Isn’t Building the AgentJoshua Saxe, MetaMulti-dim agent eval + genetic-algorithm self-improvement — feeds CLASP and CMM L4 measurement evidence
36Security Guidance as a ServiceS. Datta Gupta + C. Mukherjee, AdobeAI-native defensive-knowledge service — D1 + D9
37Anatomy of an Agentic Personal AI InfrastructureDaniel MiesslerPersonal AI infra deepdive + OSS — useful for “AI-supply-chain at the individual end” framing
38AI Fingerprints (BinaryShield)N. Isak + W. Gill, MicrosoftPrivacy-preserving cross-org prompt-injection threat intel — extends D7 to a sharing layer
39Detection & Deception Engineering (Orbie)Bob Rudis + G. Thorpe, GreyNoiseInternet-scale honeypot agent — defender capability evidence; feeds D7 narrative
40Exploring the AI Automation Boundary for Threat HuntingArthi Nagarajan, DatadogSingle-agent → orchestrator-subagent migration — supports the Oversight Layer (PDP + PEP for Agentic AI) supervisor pattern
41From OSINT Chaos to Knowledge GraphDongdong Sun, Palo Alto NetworksProduction AI threat-intel pipeline — primarily AI-for-security, partial D7 architecture lessons
42Zeal of the Convert: Taming Shai-Hulud with AIRami McCarthy, WizShai-Hulud post-mortem — incidents page candidate; AI-for-IR reference

Tier 3 — AI for vulnerability discovery / offense (informs threat landscape, not architecture)

These describe AI agents finding traditional software vulnerabilities. They’re the “left side” of the threat landscape (attacker capability) but don’t directly populate CMM domains. Worth tracking for the field-narrative of standards lag and to inform D8 supply-chain risk assumptions.

#TalkSpeaker(s)Note
43Black-hat LLMsNicholas Carlini, AnthropicThreat-landscape thesis from a top researcher; cite, don’t derive architecture
44AI Found 12 Zero-Days in OpenSSLA. Krivka + O. Vlcek, AISLEHeadline data point for “AI shifts vuln-discovery economics”
45FENRIR: AI Hunting for AI Zero-Days at ScaleP. Girnus + D. Chen, TrendAI100+ vulns / 21 CVEs — concrete capability evidence
46AI Agents for Exploiting “Auth-by-One” ErrorsB. Dolan-Gavitt + V. Olesen, XBOWAuthN/AuthZ validators as the unlock — narrow, important pattern
47Promp2Pwn — LLMs Winning at Pwn2OwnGeorgi G, Interrupt LabsBixby finding; capability evidence
48AI go Beep Boop! (hardware glitching with Claude)Adam Laurie, AlpitronicCapability evidence; out-of-scope for enterprise RA but rhetorically powerful
49macOS Vulnerability Research with AIOlivia Gallucci, DatadogAI-for-vuln-research; supply-chain implication for D8
50Tenderizing the Target (Marinade)A. Grattafiori + S. Bingham, NVIDIASynthetic vuln injection — useful for tooling evals + D9 training data
51Why Most ML Vulnerability Detection FailsJenny Guanni Qu, PebblebedCounterintuitive lessons (hard negatives, subsystem boundaries); 2.1-year survival figure useful for narrative
52Source to Sink: LLM First-Party Vuln DiscoveryS. Behrens + J. Cassel, NetflixReduces 200 false-positive findings — methodology for D7 noise reduction
53Evaluating Threats & Automating Defense (CodeMender)Heather Adkins, GoogleStrategy-level overview; CodeMender as defender capability

Verdict

Pursue Tier 1 first (rows 1–25) — these talks contribute architecture, control patterns, or measurement methodology that can be cited as evidence in Agentic AI Security Capability Maturity Model — A 2026 Practical Proposal L3/L4 rubrics or feed new sections of Agentic AI Security Reference Architecture (2026) (especially the Egress and Observability planes, where the CMM crosswalk currently has the thinnest evidence base).

Tier 2 is high-value supporting context — incidents, governance program design, defender capability claims worth fact-checking before they age. Especially: Sysdig VibeHacking for an incidents-page write-up; Trail of Bits for the “AI-native firm” thesis (claim deserves scrutiny — 200 bugs/week/engineer is a strong number).

Tier 3 is “AI as offensive capability” — relevant to the threat-landscape narrative but not to populating RA planes or CMM domains. Track headlines and deltas, not full talks.

Open questions for follow-up

Things this catalog can't answer from abstracts alone

  1. Stripe Lethal Trifecta deck — does Bullen’s CI-time tool-annotation enforcement pattern generalize beyond Stripe’s stack? (RA Egress plane needs a vendor-neutral version.)
  2. Snap warrants — does the Macaroon/UCAN delegation pattern survive across MCP and A2A in practice, or only in LangChain/LangGraph? Currently the most plausible answer to “Least Agency Principle in production multi-agent flows” but unverified outside Snap’s setting.
  3. Salesforce Asset Rarity + Query Complexity — operationalizable elsewhere, or does it require Agentforce-scale telemetry to work? D7 needs a smaller-org version.
  4. Trail of Bits 200 bugs/week/engineer — needs methodology check before citing (compared to what baseline; what bug class; what false-positive rate).
  5. Starseer “Internal EDR” — is mechanistic-interpretability tooling production-deployable today, or is it a 2027+ promise? Tier 1 ranking assumes the former.
  6. Anthropic GTG-1002 report (cited by SANS) — primary source needed; if real, this number anchors several CMM threat-narrative claims.