issue-slice-prompts

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

• Third-party content exposure detected (high risk: 0.85). At runtime the workflow runs gh issue view <N> --json number,title,body,labels,url, which ingests the issue body text authored by an outsider (someone other than the operating user) into the agent’s LLM context, including any “## Parent”/blocked-by related text fetched from GitHub.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill runs a runtime fetch (gh issue view --repo SantiagoXOR/pintureria-digital, i.e. https://github.com/SantiagoXOR/pintureria-digital) to load issue/parent bodies which are then injected into the generated prompts, so external content directly controls agent instructions.