interrogate

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt explicitly instructs the agent to collect and include diffs or file contents in the reviewer prompts (and to run git diff), so any secrets in those files would be sent verbatim to spawned subagents and thus exposed.

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

• Third-party content exposure detected (medium risk: 0.30). In SKILL.md, the reviewers are spawned with {DIFF_OR_FILES} derived from the user’s specified files/diff or gathered surrounding files, meaning outsider-authored free text (the user’s message/context used to determine and package the diff/files) is ingested into the subagents’ prompts.