plankton-code-quality
Pass
Audited by Gen Agent Trust Hub on Apr 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill configures
PostToolUseandPreToolUsehooks to automatically execute shell commands (such asnode,ruff,biome, andshfmt) whenever a file is written or edited. This automated execution pipeline is the intended mechanism for the Plankton methodology described inSKILL.mdto ensure write-time code quality enforcement. - [PROMPT_INJECTION]: The skill architecture creates a surface for indirect prompt injection by ingesting data from external linter tools and delegating autonomous fixes to subagents.
- Ingestion points: Violation data is captured from the output of third-party CLI tools (e.g.,
ruff,biome,shellcheck) through themulti-linter.jsscript. - Boundary markers: The instructions and configuration examples do not define explicit delimiters or instructions for the subagent to disregard instructions potentially embedded within linter violation messages.
- Capability inventory: Subagents are granted the capability to edit and write files to the filesystem to perform code fixes.
- Sanitization: The skill does not describe any sanitization or validation of the ingested linter output before it is passed to the subagent for processing.
Audit Metadata