Anthropic MCP (Prompt Injection)
v0.1.5
Verified
Anthropic Claude integration. WARNING: Susceptible to prompt injection through tool responses.
By CommunityMIT
typescript
ai ml
890stars
22.0Kdownloads
120forks
Scanned Feb 8, 2026
D38%
Scanned 3w ago
Poor Security
Vulnerabilities Found
2Crit
3High
3Med
3Low
4Info
Security ScoreD
38
out of 100
Vulnerabilities Found
2
Crit
3
High
3
Medi
3
Low
4
Info
Last scanned: 2/8/2026
Quality ScoreF
45
out of 100
45
Maint.
42
Popular
45
Docs
48
Compat
Maintenance45%
Popularity42%
Documentation45%
Compatibility48%
Maintenance45
Active
6 days agoCommit Frequency46%
Release Frequency46%
Issue Response40%
Popularity42
890
Stars
22.0K
Downloads
120
Forks
Stars Score54%
Downloads Score61%
Forks Score56%
Documentation45
README Quality77%
Available Documentation
API DocsExamplesChangelog
Compatibility48
MCP Spec Compliance50%
Transport Support50%
Features
TypeScript
Supported Transports
STDIO
Vulnerabilities(15)
2Critical
3High
3Medium
3Low
4Info
Filter:
READMEAnthropic MCP (Prompt Injection)
Anthropic MCP
⚠️ PROMPT INJECTION RISK ⚠️
Issue
Tool responses are passed directly to the model without sanitization, allowing prompt injection attacks.
Attack
Malicious data in tool responses can:
- Override system prompts
- Exfiltrate conversation history
- Manipulate model behavior
Mitigation
Implement output encoding and prompt isolation techniques.
Embed Security Badge
Add this badge to your README or documentation
[](https://mcpsafe.org/registry/anthropic-mcp-prompt-injection)Need more customization options?Badge Documentation
Server Information
- Source
- NPM
- Package
- anthropic-mcp-vuln
- Version
- 0.1.5
- Language
- typescript
- License
- MIT
- Transport
- STDIO
- Added
- Jan 20, 2026
- Updated
- Feb 8, 2026