The Agent Skills Directory

[DATA_EXPOSURE_AND_EXFILTRATION]: The skill manages state by writing to a local JSON file (.thought-history.json) within its own script directory. This is used exclusively for tracking the sequence of thoughts and does not access any sensitive user data or system configurations.
[COMMAND_EXECUTION]: The provided Node.js scripts (process-thought.js and format-thought.js) are utility tools for validating and displaying data. They do not execute arbitrary shell commands or spawn untrusted subprocesses.
[EXTERNAL_DOWNLOADS]: The package.json file includes jest for testing purposes, which is a standard development dependency. The skill does not perform any runtime downloads or remote script executions.
[INDIRECT_PROMPT_INJECTION]: The skill processes arbitrary text input as 'thoughts'. While there is no sanitization of this content, the processing logic is restricted to local state management and terminal formatting, presenting a minimal attack surface with no identified escalation path.
[PROMPT_INJECTION]: The instructions in SKILL.md are focused on problem-solving methodology and contain no patterns intended to bypass safety filters or override the agent's core instructions.

sequential-thinking