session-report

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill requires reading /tmp/session-report.json and embedding its full JSON verbatim into the output HTML (and filling report fields from it), so any secrets, API keys, or sensitive prompt content inside that JSON would be output exactly as-is.

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

• Third-party content exposure detected (high risk: 0.85). The skill runs node analyze-sessions.mjs --json ... which reads outsider-authored transcript text from ~/.claude/projects/**.jsonl (and related .meta.json) and embeds it into the HTML as readable JSON (#report-data), including top_prompts[].text and cache_breaks[].context derived from those transcripts.