plugin-validator

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill explicitly searches for hardcoded secrets (SKILL.md:31) and requires a validation report listing the "specific issue" and ready-to-run fix command (SKILL.md:46), which together imply the agent may need to include offending secret values verbatim in its output.

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

• Third-party content exposure detected (medium risk: 0.65). The skill’s required workflow validates a target plugin by reading outsider-authored files from the repository/plugin directory at runtime (e.g., README.md, plugin.json, marketplace.extended.json, and component frontmatter under commands/, agents/, skills/), and those file contents can be surfaced in the produced validation report that is fed back into the agent/LLM context (indirect prompt injection risk from arbitrary plugin markdown/YAML/JSON text).