The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script located at scripts/analyze.py. This script is used to aggregate and analyze usage statistics from the user's local environment. This is the intended behavior of the skill and is triggered via the agent's shell execution capabilities.
[DATA_EXPOSURE]: The analysis script reads interaction history from ~/.claude/history.jsonl and project-specific session files in ~/.claude/projects/. While these files contain sensitive interaction data, the processing is performed locally, and the results are presented directly to the user in the chat interface as a summary.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted historical data (past user prompts). There is a theoretical surface where malicious instructions previously entered into the history could be re-processed during the analysis phase. However, the skill instructions provide a strict 5-section report format which acts as a structural constraint for the agent's output.

perf-review