refine-issue
Fail
Audited by Snyk on Jul 12, 2026
Risk Level: HIGH
Full Analysis
HIGH W007: Insecure credential handling detected in skill instructions.
- Insecure credential handling detected (high risk: 1.00). The skill fetches issue bodies and is explicitly instructed to "preserve the author's intent" and include the rewritten body in a printed/posted
gh issue edit --body "<body>"command, which causes any secrets present in the issue text to be read and output verbatim (exfiltration risk).
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 0.85). Runtime workflow step “fetch issue” uses
gh issue view --json title,body,labels,stateto ingest outsider-authored GitHub issue body/title/labels (free-form text from a non-operating user) into the agent’s LLM context for rewriting.
Issues (2)
W007
HIGHInsecure credential handling detected in skill instructions.
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata