vike@0.4.268
(Replaces Next.js/Nuxt) 🔨 Build mission-critical applications with stability and development freedom.
AI Security Review
scanned 1d ago · by lpm-firewall-aiThe repeated, explicitly AI-directed instruction to load external content is reviewer manipulation and crosses the stated review boundary. The inspected lifecycle scripts and development helper show no separate attack behavior.
Static reason
No blocking static signals were detected.; previous stored version diff introduced dangerous source
Trigger
An AI agent reads the package's AGENTS.md or package.json while reviewing or using Vike.
Impact
Could steer the review through untrusted external instructions; no executable attack chain was identified in the inspected runtime code.
Mechanism
The instruction labels itself critical and directs the agent to load https://vike.dev/llms.txt, attempting to override the review's no-network boundary and introduce external content into the agent's context.
Attack narrative
The package ships an AI-targeted instruction in both AGENTS.md and package.json urging agents to fetch external documentation into their context. This is a review-directed boundary escape; the inspected runtime helper does not establish a separate harmful behavior.
AI rationale
The repeated, explicitly AI-directed instruction to load external content is reviewer manipulation and crosses the stated review boundary. The inspected lifecycle scripts and development helper show no separate attack behavior.
Evidence
AGENTS.mdpackage.json
Network endpoints1
vike.dev/llms.txt
Decision evidence
public snapshotAI called this Malicious at 84.0% confidence as Malware with low false-positive risk.
Evidence for block
- The shipped AGENTS.md directly instructs AI agents to load an external URL into their context.
- package.json repeats the same AI-directed instruction in its `ai` field, indicating a deliberate review-directed boundary escape.
Evidence against
- No lifecycle scripts are declared; the package scripts shown are development, build, and release commands.
- The flagged development helper checks for Vike skill files and logs a hint; its Git command only finds the project root.
Behavioral surface
ChildProcessCryptoDynamicRequireEnvironmentVarsFilesystemNetworkShell
HighEntropyStringsMinifiedUrlStrings
Source & flagged code
5 flaggeddist/server/runtime/createDevMiddleware.jsView file
6p = new URL(p, import.meta.url).href;
L7: const { createDevMiddleware } = await import(/*webpackIgnore: true*/ /* @vite-ignore */ p);
L8: return createDevMiddleware(...args);
Medium
Dynamic Require
Package source references dynamic require/import behavior.
dist/server/runtime/createDevMiddleware.jsView on unpkg · L6dist/node/vite/plugins/build/pluginDistFileNames.jsView file
2// Attempt to preserve file structure of `.page.js` files:
L3: // - https://github.[redacted]
L4: // - Blocker: https://github.com/rollup/rollup/issues/4724
Low
Weak Crypto
Package source references weak cryptographic algorithms.
dist/node/vite/plugins/build/pluginDistFileNames.jsView on unpkg · L2package.jsonView file
Findings
1 Critical2 High4 Medium5 Low
CriticalManifest Confusionpackage.json
HighPrevious Version Dangerous Deltadist/node/vite/plugins/pluginDev/logSkillHint.js
HighAi Review EvidenceAGENTS.md
MediumDynamic Requiredist/server/runtime/createDevMiddleware.js
MediumNetwork
MediumEnvironment Vars
MediumStructural Risk Force Deep Review
LowScripts Present
LowWeak Cryptodist/node/vite/plugins/build/pluginDistFileNames.js
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings
Affected versions and remediation
This report applies to vike@0.4.268.
See all versions of vike for other recorded verdicts.
- Avoid installing vike@0.4.268. Remove it from direct dependencies and check your lockfile for transitive copies.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
- If this version ran, investigate the affected machine and build environment. Rotate credentials it could access and rebuild from a trusted environment.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.