This video could not be loaded.
Recording
Academy Meetup: Software Bills of Materials in Practice
Oct 1, 2026, 11:30 AM
AMAs, deep-dives, release walkthroughs, and course sessions — something live every week.
A live walkthrough of the first twenty four hours of a zero-day event: answering whether you are affected in minutes, blocking with the package firewall, mitigating before a patch exists, and communicating with customers. First Thursday, 5:00 PM IST, 60 minutes, on Google Meet.
A live session on replacing manual evidence collection with automation across System and Organization Controls 2 (SOC 2), ISO/IEC 27001 and regional frameworks, and on keeping one control set mapped to many. First Thursday, 5:00 PM IST, 60 minutes, on Google Meet.
A live session on securing AI assisted development: assistant and agent inventory, Model Context Protocol server governance, prompt injection defence and agent identity. First Thursday, 5:00 PM IST, 60 minutes, on Google Meet.
A live session on cloud and cluster posture: infrastructure as code checks before deployment, drift between code and cluster, admission control and prioritising posture findings against reachable vulnerabilities. First Thursday, 5:00 PM IST, 60 minutes, on Google Meet.
A live session on measurement that survives scrutiny: risk score, coverage, mean time to remediate, service level breaches and exposure to actively exploited vulnerabilities, and how to report them to a board. First Thursday, 5:00 PM IST, 60 minutes, on Google Meet.
Where the Safeguard team shows up in person — the same conference calendar as safeguard.sh/company/events.
Talk: "Coordinated disclosure pipelines for zero-day candidates surfaced by reachability plus LLM hypothesis."
Live demo of self-healing container pipelines with Griffin auto-fix.
Workshop in the AI Village on adversarial prompt-injection eval methodology.
Briefing on the security-only training corpus thesis and the Griffin attention pattern.
Partner-track demo of reachability-aware auto-fix in an AWS-native CI/CD pipeline.
Live Griffin reasoning demo against a held-out CVE set. Talk: "Engine plus LLM beats pure-LLM bug hunters."
Lightning talk on MCP-server governance and sensitive-data egress guardrails. The open-source policy templates library was released after the event in response to attendee feedback.
Workshop on integrating reachability analysis into existing SAST + SCA tool chains. Strong interest from financial services attendees on DORA evidence pipelines.
Panel + booth focused on AI for security: the security-only training corpus thesis, MCP-server governance, and the practical economics of running reasoning models inside regulated environments. Strong engagement from Indian financial services and public sector attendees.
Community-track talks on reachability + LLM bug hunting and on coordinated disclosure workflows. Hands-on workshop with the Safeguard CLI and the policy templates library.
Return engagement with deeper coverage of cross-package taint chain reasoning and Griffin XL's 256k-context retrieval gates. Live demo against a held-out exploit set.
Two talks: one on the Griffin reasoning trace contract and one on adversarial robustness in security-tuned LLMs. Slides, demo notebooks, and the eval-set methodology published openly.
Briefing on the engine-plus-LLM thesis for software supply chain security. Heavy interest in the structured trace contract and the adversarial disproof pass. Slides published on the blog.
Booth + speaking. Lit up the early Griffin reasoning trace demo on stage and showed the held-out CVE evaluation methodology. First public unveiling of Eagle as a standalone ranking model.