Deep dives, practical guides, and incident analyses from engineers who build Safeguard. No fluff, no vendor FUD — just what you need to ship secure software.
When LLMs review PRs, triage issues, and fix builds, every commit message becomes attacker input. The concrete attack paths through GitHub Actions and what blocks them.
Pure-LLM vulnerability scanners hit production around 2024. By 2026 their failure modes are documented. Reachability remains the backbone — and the LLM is most useful on top of it.
Some tool calls cannot be undone. Out-of-band confirmation is the cheapest defense for that small set, and the most expensive thing to skip.
An attacker who can swap the model behind an API call can read every prompt and shape every response. The emerging trend in 2026 is model substitution as an attack class with its own techniques and disclosures.
A threat model for sandbox escapes in Model Context Protocol servers, mapping attack surfaces from tool execution environments to host processes and shared state.
A senior engineer's guide to training data poisoning defenses in 2026, from split-learning detection to provenance attestation and continuous pipeline monitoring.
Cody's codebase-wide context is valuable for security review. Griffin AI adds reachability, taint, and policy grounding that Cody doesn't target.
A hijacked tool call is more consequential than a hijacked response. The defence requires the tool layer to police the model, not the other way around.
A one-hour cycle from vulnerability finding to merged fix is achievable in 2026, but only with a pipeline designed for it. Here is what that pipeline looks like.
Weekly insights on software supply chain security, delivered to your inbox.
Your first fix PR is minutes away.
No sales call required, even your agent can complete the purchase over MCP.
