deserialization
Safeguard articles tagged "deserialization" — guides, analysis, and best practices for software supply chain and application security.
65 articles
Python code injection: eval, exec, and pickle explained
eval(), exec(), and pickle.load() can each hand an attacker a Python interpreter — CVE-2020-1747 shows how one unsafe deserialization call became a real RCE.
CTF writeup patterns: serialization and cryptographic puzzles, decoded
CVE-2013-0156 let attackers RCE Rails by feeding YAML into a parameter parser — the same insecure-deserialization pattern CTF players train on every weekend.
Log4Shell and Spring4Shell, years later: why the same bug keeps coming back
CVE-2021-44228 scored a perfect CVSS 10.0 and hit CISA's Known Exploited Vulnerabilities list the day it was published — the root cause hasn't gone away.
PHP code security fundamentals: injection, deserialization, and file inclusion
PHP still powers over 70% of server-side websites, and its three oldest vulnerability classes — injection, deserialization, and file inclusion — remain the most common findings in 2026.
CVE-2022-1471: Inside the SnakeYaml Deserialization RCE
CVE-2022-1471 scored 9.8 CRITICAL under NIST's CVSS calculation — a single YAML tag could hand attackers remote code execution in any Java app parsing untrusted input.
The most common Spring Boot security misconfigurations, and how to fix them
CVE-2026-40976 let anonymous users hit /actuator/env and /actuator/heapdump on default Spring Boot 4 filter chains, CVSS 9.1 — here's how to actually harden Spring Boot.
Unsafe Deserialization in Swift: NSCoding, Codable, and Safer Patterns
Two 2019 iOS zero-click bugs, CVE-2019-8646 and CVE-2019-8647, both traced back to NSKeyedUnarchiver — a reminder that Swift's Objective-C legacy still hides deserialization risk.
How task-scheduler RCEs become cryptomining botnets
Two chained Apache Airflow CVEs and a Rundeck YAML deserialization bug show how scheduler tools turn one flaw into unauthenticated RCE and persistent mining.
The Python pickle security model, explained
Python's own docs warn that unpickling can execute arbitrary code — yet pickle is still the default weight format behind millions of ML model downloads.
Elixir and Phoenix Security Best Practices: BEAM Footguns and the Hex Supply Chain
Phoenix is safe by default, but the BEAM has its own footguns — binary_to_term, atom exhaustion, dynamic eval — and the Erlang/OTP runtime beneath it shipped a CVSS 10.0 pre-auth SSH RCE in 2025.
OWASP A08: Software and Data Integrity Failures — A Deep-Dive Guide
Software and Data Integrity Failures rank #8 in the OWASP Top 10 (2021). A deep dive into insecure deserialization, unsigned updates, SolarWinds, and real CVEs.
Pickle and Deserialization Security in Python
Unpickling untrusted data is arbitrary code execution, by design. Here is why pickle is dangerous, where it hides, and what to use instead.
Self-healing security runs on Safeguard.
Your first fix PR is minutes away.
No sales call required, even your agent can complete the purchase over MCP.