The Agent Skills Directory

[SAFE]: The skill implements persistent memory management using a standard file-based approach. It creates and maintains a local directory structure (.agent-memory/) to store session learnings, decisions, and project context.
[DATA_EXPOSURE]: The build operation scans common project configuration files (such as package.json, pyproject.toml, and .env.example) to bootstrap the memory system. This is limited to publicly visible project documentation and does not attempt to read sensitive files like .ssh/ or active .env files.
[PROMPT_INJECTION]: By processing existing project documentation to generate memory files, the skill has an indirect prompt injection surface. However, the instructions emphasize distillation and summarization rather than direct execution of content found in the documents, which is a standard safety practice for this use case.

agent-memory