PPhying中文
Curated security research, vulnerabilities, advisories and tools for practitioners.

Fickling’s new AI/ML pickle file scanner

blog.trailofbits.com | tool | #ai-security | #open-source | #machine-learning | #static-analysis | #tool-release | #supply-chain-security | #trail-of-bits | #fickling | #pickle
📄 Raw .md⬇️ Download .md🔗 Original ↗

Summary

Trail of Bits updated Fickling, its pickle file scanner, with an allowlist approach that blocks malicious imports in AI/ML environments, countering supply-chain attacks hidden inside model files.
Published
Collected

Related coverage

tool · github.com
Droid ASC: High-Performance On-Demand Android Decompiler Front-End for Mobile Security & AI Agents
Radically reduces Android reverse-engineering overhead from tens of minutes and gigabytes of memory to milliseconds and megabytes, providing an ideal, stateless code analysis primitive for autonomous security agents and vulnerability researchers.
tool · certik.com
CertiK Skill Scanner: The Antivirus for the AI Age
CertiK Skill Scanner flags execution-stage risks in third-party AI Skills—data exfiltration, shell execution, file misuse—scoring repos, URLs, or ZIPs 0-100 with up to 90.5% precision.
blog · blog.trailofbits.com
Auditing in the age of (good enough) AI
Trail of Bits explains how agents built tooling and a Lean model for the Miden zkVM audit, finding a flaw enabling Falcon signature forgery and yielding 95 machine-checked proofs.
blog · blog.trailofbits.com
1Password's AI patching benchmark is misleading
Trail of Bits says 1Password's FLAWED benchmark misstates AI patch quality, mixing wrong-fix prompts and no-test trials into a 26% clean-fix headline, and ships two agent skills for better fixes.
blog · blog.trailofbits.com
Rust-proof your code with our new Testing Handbook chapter
Trail of Bits added a Rust testing chapter to its Testing Handbook: Miri, property testing, Clippy, zeroization, Kani model checking and dependency vetting, plus the rust-review Claude Code plugin.
research · certik.com
Skill Scanning Is Not a Security Boundary
This article argues that scanning third-party OpenClaw skills (e.g. Clawhub's VirusTotal and AI review pipeline) is not a security boundary, since skills execute with full privileged-runtime access.
Back