issue-to-pr

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The preflight step explicitly says it prints KEY=value pairs and tells the agent to "substitute the values you read" into subsequent shell calls, which requires the LLM to insert literal values (potentially secrets/tokens) verbatim into generated commands/outputs.

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

• Third-party content exposure detected (medium risk: 0.30). The workflow reads outsider-authored GitHub content at runtime via triage-evidence.sh <N> / review-check.sh <N> (Step 2 and Step 11), which ingest the issue/PR text and review decisions tied to user-submitted content.