verasic-secbot

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill instructs the agent to execute reviews in-conversation and to relay subagent/scanner reports verbatim (with only limited stripping), which can force the LLM to include detected secrets or raw credential strings verbatim in its output — an exfiltration risk.

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.90). These links include direct raw GitHub shell scripts (curl|bash style installers) from individual repos (including an external opengrep installer) which are high-risk to execute without review because they can run arbitrary code; the SECURITY.md blob is benign documentation.

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

• Third-party content exposure detected (high risk: 0.85). Outsider-authored free text can enter the LLM context via the runtime “Diff” content and full changed file contents (branch/uncommitted/untracked sources), which are not authored by the operating user but are read and relayed to the verasic-secbot-reviewer subagent for the LLM pass.