taste

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly requires recording full badge markdown including raw URLs with all query parameters and embedding sampled file snippets into the generated TASTE.md/TASTE.html output, which forces the LLM to include any secret tokens or sensitive query params found in repo files verbatim.

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

• Third-party content exposure detected (medium risk: 0.30). Workflow step 4 reads and ingests arbitrary file contents from outsider-authored Git repositories provided as arguments (e.g., README.md, CLAUDE.md/AGENTS.md/CODEX.md, and prioritized implementation/config/CI/test files after cloning), which can contain indirect prompt-injection text.