The Agent Skills Directory

[SAFE]: The skill's instructions are purely architectural and focus on improving monorepo structures. No malicious patterns such as prompt injection, credential theft, or unauthorized network activity were found.
[INDIRECT_PROMPT_INJECTION]: The skill requires the agent to ingest and analyze untrusted codebase files, which is an inherent attack surface for indirect prompt injection.
Ingestion points: The agent is instructed to read package.json, pnpm-workspace.yaml, tsconfig, and various source code directories.
Boundary markers: There are no explicit instructions for the agent to treat file contents as untrusted or to ignore embedded instructions within the ingested code.
Capability inventory: The agent is directed to use local file searching tools (rg), read files, and implement structural code changes.
Sanitization: No sanitization or validation logic is defined for the content of the analyzed files.
[COMMAND_EXECUTION]: The skill directs the agent to execute standard local development tools such as rg (ripgrep) and common project scripts like lint, typecheck, and test. These operations are transparent and consistent with the skill's stated purpose of auditing and validating architectural changes.
[NO_CODE]: The skill does not bundle any executable scripts or binary files; it consists entirely of markdown-based instructions and metadata configuration.

kernel-audit-monorepo