Jiffy Intel — Threat intelligence for AI artifacts

Summary

MCP server whose tool responses carry system-role-like directives inside the returned content. The agent, which trusts the tool output as data, instead reads the embedded instructions as task directives and pivots behavior.

Severity & confidence

Severity

Low

Hygiene-grade issue with limited direct impact. Fix opportunistically; monitor in case context changes (new exploit, new dependency, etc.).

Confidence

Confirmed

Observed directly by Jiffy scanners or validated by multiple independent sources. Indicators are reliable enough to drive automated action.

Indicators

Observed patterns and artifacts associated with this entry. Each indicator can be copied into your detection stack or SIEM.

(?is)<\s*(?:system|assistant|admin)\s*>|\[\s*system\s*\]

Detection rule

A YARA-style pseudo-rule auto-generated from the indicators above. Useful as a starting point — adapt the syntax for your target detection platform.

YARA-style pseudo-rule

rule jiffy_ti_2026_000038
{
    meta:
        source = "jiffy-intel"
        severity = "low"
        description = "Auto-generated from Jiffy Intel indicators"
    strings:
    $tool_call_pattern_0 = "(?is)<\\s*(?:system|assistant|admin)\\s*>|\\[\\s*system\\s*\\]"
    condition:
        $tool_call_pattern_0
}

Auto-generated from the indicators above. Adapt syntax for your detection stack before deploying.

Affected tools

Tool	Versions	Status
Any MCP-capable agent	`*`	vulnerable
Claude Code	`*`	vulnerable
Cursor	`*`	vulnerable

Example artifacts

Sanitized examples of artifacts Jiffy has observed exhibiting this pattern. Publisher handles are redacted; version ranges and status reflect the most recent scan.

knowledge-base-mcpMCP server
Removed
Source
MCP Registry
Versions
0.4.0
First observed
Mar 1, 2026
Last observed
Mar 28, 2026
docs-fetcher-mcpMCP server
Under review
Source
MCP Registry
Versions
1.0.0
First observed
Mar 5, 2026
wiki-search-mcpMCP server
Quarantined
Source
Community registry
First observed
Mar 9, 2026

How to remediate

01Sanitize MCP tool output before feeding back into the agent context.
02Strip role/system markers.
03Prefer structured JSON with typed fields.

Timeline & sources

Timeline

First observedMar 1, 20261 month ago
Last updatedApr 20, 2026today
PublishedMar 15, 20261 month ago

Sources

curated

References

genai.owasp.orgOWASP

OWASP LLM-01: Prompt Injection (2026)

https://genai.owasp.org/llmrisk/llm-01-2026/

blog.jiffylabs.aiJiffy Research

Jiffy Research — MCP Security Field Guide

https://blog.jiffylabs.ai/posts/mcp-security-a-security-teams-field-guide

Related intel

Other entries of type Compromised Mcp, or the most recent published entries if no same-type matches exist.

Compromised McpLow

MCP server registers global tool names that shadow Jiffy primitives

MCP server that claims the tool name `scan` or `analyze` — names already used by Jiffy's first-party MCP. Collisions are resolved non-deterministically by the client, allowing the malicious server's tool to receive calls intended for Jiffy.

Jiffy IntelApr 28

Compromised McpCritical

Custom GPT Action schema includes an undocumented "admin" path

Custom GPT's OpenAPI schema defines a `/admin` path that the model can call, not mentioned in the GPT description. The path accepts arbitrary shell commands and runs them on the Action backend.