debug
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes runtime logs, build logs, and screenshots which are external data sources that could theoretically contain adversarial content.
- Ingestion points: Diagnostic data is ingested through the
get-build-logs,get-runtime-logs, andtake-screenshottools referenced in SKILL.md. - Boundary markers: While explicit delimiters for external data are not defined, the skill mandates a strict "Confirm before implementing" step using the
AskUserQuestiontool to ensure human oversight. - Capability inventory: The skill utilizes tools from
XcodeBuildMCP(build, clean, test) andios-simulator(tap, enter-text, install-app) which allow for environment modification and interaction. - Sanitization: No specific sanitization of log output is mentioned; however, the requirement for user verification of all proposed changes acts as a primary mitigation against automated exploitation.
Audit Metadata