The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes several standard shell utilities including ps, awk, jq, grep, and stat to monitor system process memory and analyze heap snapshots. These commands are used for legitimate profiling purposes and operate on local data within the project directory.
[DATA_EXPOSURE]: The skill inspects local project files such as package.json, requirements.txt, and pyproject.toml to detect the runtime environment and installed profiling tools. It also creates a local cache in .claude/cache/memory-leak/ to optimize subsequent runs. No evidence of sensitive data harvesting or remote exfiltration was found.
[PROMPT_INJECTION]: The skill performs static analysis on source code using grep to find event listeners and timers. While this involves processing user-controlled data, the risk of indirect prompt injection is low as the skill extracts specific matches rather than interpreting arbitrary text as instructions.
[REMOTE_CODE_EXECUTION]: The skill generates script snippets for memory profiling (e.g., Node.js heap snapshot code) and provides instructions on how to run them. This code generation is based on fixed templates and does not incorporate untrusted remote content.

memory-leak