The Agent Skills Directory

[DATA_EXFILTRATION]: The skill reads from ~/.claude/projects/, which contains the complete history of user interactions with Claude Code. This conversation history is sensitive and may contain private information, code, or context from previous sessions.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes historical chat logs as input data. If these logs contain malicious instructions, they could potentially influence the agent's behavior when it later reads the organized markdown files.\n
Ingestion points: Reads .jsonl session files from the ~/.claude/projects/ directory.\n
Boundary markers: Extracted content is placed directly into markdown files under headers without specific delimiters or instructions to ignore embedded commands.\n
Capability inventory: The skill can create directories (mkdir), move files (mv), and perform Git operations (git add, git commit).\n
Sanitization: Filters out system commands and tool outputs using jq and grep, but preserves raw user input strings without escaping or validation.

chat-history