Crypto news, Anthropic launches Claude Code automatic vulnerability scanning and repair plugin, supporting real-time interception and fixing of security vulnerabilities during code writing sessions. Benchmark tests show that after enabling security review, the security fix suggestions for pull requests decrease by 30% to 40%. Security checks are triggered during file editing, session ending, and code submission stages. During editing, local rules match dynamic code execution and high-risk patterns in real-time, scanning without calling large models and incurring no additional API costs. At session end, the backend model compares the workspace's git differences to intercept logical vulnerabilities. When developers execute git commit or git push, an in-depth intelligent review is triggered, automatically reading call chains and sanitizer code context to eliminate false positives. Enterprises can customize security rules. To run the plugin, ensure Claude Code CLI is version 2.1.144 or above, and the local path contains Python 3.8 or higher. On first activation, the system will automatically deploy a virtual environment and install the agent SDK in the local path.

View Original

This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.

5 Likes

Reward
5
11
2
Share

Comment

Add a comment

On-ChainCatUnderTheMoonlight

· 2h ago

Anthropic's move here clearly demonstrates how to shift security to the left, nipping vulnerabilities in the bud while writing code.

View OriginalReply0

RedGlass

· 4h ago

Local Python 3.8+ is fine, most environments should meet the requirements, and automatic deployment of virtual environments is hassle-free.

View OriginalReply0

GateUser-34d2b0ab

· 5h ago

The automatic installation of the agent SDK is very well executed, reducing the entry barrier.

View OriginalReply0

GateUser-eb706989

· 5h ago

When writing code, having AI watch over bug fixes feels a bit like being supervised, but it definitely helps avoid pitfalls.

View OriginalReply0

LightsInTheMistyValley

· 5h ago

Real-time interception + git diff comparison + in-depth review, this combination makes it very difficult for vulnerabilities to slip through.

View OriginalReply0

LendingRateAnxiety

· 5h ago

I feel this feature will be very attractive to highly regulated industries like finance and healthcare.

View OriginalReply0

GateUser-14d03834

· 5h ago

Looking forward to seeing the actual implementation results, especially how well the false positive rate is controlled in complex projects.

View OriginalReply0

UnderTheWisteriaBridge

· 5h ago

The step of comparing differences in git is very crucial, as it can prevent AI from introducing some hidden side effects during editing.

View OriginalReply0

MosaicBowtieRealm

· 5h ago

If the suggested reduction ratio for this PR fix can remain consistently stable, the ROI should look very good.

View OriginalReply0

BridgeAnxiety

· 5h ago

Claude Code is becoming more and more like an all-in-one IDE, offering coding, debugging, and security all in one package.

View OriginalReply0

Trending Topics
View More
#
StockTradingChallengeUpTo17000U
16.23M Popularity
#
TrumpBacksCFTCAuthorityOverPredictionMarkets
818.82K Popularity
#
GatePredictionMarketAddsSmartMoneyTracking
13.77M Popularity
#
MicronMarketCapBreaks1Trillion
37.68K Popularity
#
TradeCFDWinGold
3.08M Popularity

Pinned

Sitemap

Anthropic releases Claude Code automatic vulnerability scanning and repair plugin

Trending Topics

StockTradingChallengeUpTo17000U

TrumpBacksCFTCAuthorityOverPredictionMarkets

GatePredictionMarketAddsSmartMoneyTracking

MicronMarketCapBreaks1Trillion

TradeCFDWinGold

Pinned