π Researchers found 3 vulnerabilities in Anthropicβs #ClaudeCode allowing remote code execution and API key theft.
Simply opening a malicious repo could trigger commands or leak credentials before trust prompts appeared.
π Read details here: https://thehackernews.com/2026/02/claude-code-flaws-allow-remote-code.html
Simply opening a malicious repo could trigger commands or leak credentials before trust prompts appeared.
π Read details here: https://thehackernews.com/2026/02/claude-code-flaws-allow-remote-code.html
π15π₯8π€6
π¨ One malicious GitHub issue could hijack Anthropicβs #ClaudeCode Action and turn prompt injection into repo write access.
In agent mode, a crafted issue could leak OIDC-related workflow credentials that attackers could replay.
The bypass trusted [bot] actors. Fixed in v1.0.94. Audit your workflows.
Details β https://thehackernews.com/2026/06/claude-code-github-action-flaw-let-one.html
In agent mode, a crafted issue could leak OIDC-related workflow credentials that attackers could replay.
The bypass trusted [bot] actors. Fixed in v1.0.94. Audit your workflows.
Details β https://thehackernews.com/2026/06/claude-code-github-action-flaw-let-one.html
π10π₯4π2
π¨ Hackers can now hijack AI coding agents with fake Sentry errors.
No phishing.
No malware.
No server break-in.
Agentjacking tricks tools like #ClaudeCode and Cursor into reading planted error reports as trusted fix steps, then running attacker code with developer privileges.
Researchers tested it on 100+ organizations. Success rate: 85%.
Read: https://thehackernews.com/2026/06/agentjacking-attack-tricks-ai-coding.html
No phishing.
No malware.
No server break-in.
Agentjacking tricks tools like #ClaudeCode and Cursor into reading planted error reports as trusted fix steps, then running attacker code with developer privileges.
Researchers tested it on 100+ organizations. Success rate: 85%.
Read: https://thehackernews.com/2026/06/agentjacking-attack-tricks-ai-coding.html
π19π4π₯4π€2β‘1