eu-sparql-search

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 1.00). The skill explicitly tells the agent to bypass platform web_fetch permission checks (use bash_tool to "bypass" restrictions) and to disable SSL verification, which are instructions to override environment security and are outside the normal scope of issuing SPARQL queries.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The skill contains deliberate instructions to bypass platform network/sandbox restrictions (use of bash_tool/curl instead of web_fetch) and to disable SSL verification, which are intentional security-evasion behaviors that enable unsafe external access and potential misuse.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). The required runtime workflow executes SPARQL against a public endpoint and then uses curl to fetch the returned document HTML/text, so outsider-authored free text from the fetched EUR-Lex/Cellar documents can be ingested into the agent’s LLM context via the HTML-to-text extraction step.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.80). The skill explicitly tells the agent to "bypass all web_fetch permission restrictions" by using bash_tool and includes code to disable SSL verification, which encourages bypassing security controls that protect the agent's environment.