The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/main.cjs file uses child_process.spawn to execute a local Python script located at .claude/tools/optimization/sequential-thinking/executor.py. This is the intended execution method for the skill's standalone mode.
[EXTERNAL_DOWNLOADS]: The documentation references official packages from the @modelcontextprotocol scope via npx and the mcp Python library via pip. These are well-known and trusted resources for implementing MCP-based functionality.
[PROMPT_INJECTION]: The skill defines 'Iron Laws' and a 'Memory Protocol' which are behavioral instructions for the AI agent. These are designed to enforce a systematic problem-solving process and do not attempt to bypass safety guidelines or extract system prompts.
[DATA_EXFILTRATION]: No evidence of hardcoded credentials or unauthorized data transmission was found. The skill interacts with local project memory files (e.g., learnings.md, issues.md) to maintain state across sessions, which is standard behavior for this type of tool.
[REMOTE_CODE_EXECUTION]: The skill relies on local script execution and verified MCP packages. There are no patterns of fetching and piping remote code from untrusted sources.

sequential-thinking