codebase-onboarding

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill explicitly commands reading configuration files (e.g., cat src/main/resources/application.yml) and then generates artifacts that incorporate "actual" config-derived commands and request flows, which creates a high risk the model will output secrets (API keys, DB passwords) verbatim.

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

• Third-party content exposure detected (medium risk: 0.30). Source code is read from the user’s target repository via shell commands like cat pom.xml, cat package.json, and various find/cat/grep steps, so outsider-authored text in the repo (e.g., poisoned config/source files) is ingested to generate artifacts.