technical-debt-auditor

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The scanner explicitly returns a "secrets_found" list and requires the agent to "flag" detected secrets to the user without instructing redaction, which implies the LLM may read and relay verbatim secret values (high exfiltration risk).

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

• Third-party content exposure detected (high risk: 0.70). The Scanner explicitly runs pip-audit / npm audit / pip list --outdated / npm outdated and similar commands (see sub-agents/scanner.md and references/scoring-rubric.md) which fetch vulnerability and package metadata from public registries (PyPI/NPM/OSV) and those results are consumed by the Analyzer to compute scores and apply the security blocker, so untrusted third‑party content is ingested and can materially influence decisions.