play-billing-library-version-upgrade
Warn
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: MEDIUMMETADATA_POISONINGINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [METADATA_POISONING]: The skill's YAML frontmatter identifies the author as "Google LLC", which contradicts the provided skill author "deonmenezes". This deceptive metadata may lead users to believe the skill is an official Google tool, potentially resulting in misplaced trust regarding its ability to modify source code and execute build scripts.
- [INDIRECT_PROMPT_INJECTION]: The skill operates on untrusted project data, including source code and build files (build.gradle, libs.versions.toml). It uses this data as context and executes it via the Gradle wrapper (./gradlew). This represents a significant attack surface where malicious code within a project could influence the agent's behavior or leverage its execution capabilities.
- Ingestion points: SKILL.md Phase 1 (Discovery) instructions to locate dependency files and scan source code.
- Boundary markers: The skill does not implement specific delimiters or safety warnings for the data it processes.
- Capability inventory: The skill calls subprocesses for assembly, testing, and building in Phase 1 and Phase 3.
- Sanitization: There is no evidence of sanitization or validation of the ingested project content.
- [DYNAMIC_EXECUTION]: In references/version-checklist.md and SKILL.md Phase 2, the skill instructs the agent to "synthesize instructions" and "synthesize a new checklist" by reviewing external release notes for future versions. This requirement for the agent to dynamically generate its own logic based on live, external web content introduces risks associated with uncontrolled behavior and the processing of potentially untrusted remote data at runtime.
Audit Metadata