hunt-deserialization

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

• Third-party content exposure detected (low risk: 0.10). The skill's frontmatter lists hackerone_public as a source, which is public, outsider-authored report content, but the required workflow never directs the agent to fetch or read it, so the agent would have to actively search for those reports, which makes this low.

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

• Potentially malicious external URL detected (medium risk: 0.30). Inbound setup fetches of code that is then executed (ysoserial JAR, phpggc and JNDI-Exploit-Kit repositories) come from third-party publishers without strong first-party provenance, so the highest tier is common/reputable unofficial, which is medium.