dart-package-maintenance
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from local repository files which is subsequently used to construct and execute shell commands.\n
- Ingestion points: Reads
pubspec.yamlandCHANGELOG.mdto extract version strings and changelog entries (SKILL.md).\n - Boundary markers: The instructions do not define delimiters or provide warnings to ignore potential instructions embedded within the processed configuration or markdown files.\n
- Capability inventory: The skill utilizes impactful shell-based capabilities including
git tag,git push --tags,gh pr checkout, anddart pub publish.\n - Sanitization: There is no requirement or guidance for the agent to sanitize or validate strings extracted from the repository files before interpolating them into shell command templates.\n- [COMMAND_EXECUTION]: The skill guides the agent to execute shell commands with significant side effects.\n
- Evidence: Specifically mentions the use of
dart pub publish(orflutter pub publish) to upload packages andgit push --tagsto modify remote repository state. These commands involve network interaction and state changes based on the agent's interpretation of project files.
Audit Metadata