The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses standard development tools and commands (ls, cat, grep, npx, psql) to gather metadata about the project structure and performance metrics. These commands are executed within the context of a performance audit and are restricted to the local workspace.
[REMOTE_CODE_EXECUTION]: While the skill invokes npx for tools like @next/bundle-analyzer and lighthouse, these are well-known, industry-standard performance tools used as intended by the skill's primary purpose. No arbitrary or unknown remote scripts are executed.
[DATA_EXFILTRATION]: The skill is designed for read-only analysis. The output is a report written to a local file (.agents/memory/performance-audit-YYYY-MM-DD.md) within the workspace. No network operations or external data transmissions were detected.
[PRIVILEGE_ESCALATION]: All operations are performed with standard user permissions. There are no attempts to use sudo or modify system-level configurations.
[PROMPT_INJECTION]: The skill's instructions focus on structured orchestration and do not contain patterns aimed at overriding agent safety guidelines or hijacking behavior.

workspace-performance-audit