The Agent Skills Directory

[COMMAND_EXECUTION]: The run_codememory.sh and health_check.sh scripts manage the execution of the codememory CLI binary. The run_codememory.sh script implements a whitelist of allowed subcommands (init, status, index, watch, serve, search) to prevent the execution of unauthorized tool features. Commands are executed with configurable timeouts and retries, using standard shell quoting to handle repository paths and arguments safely.
[DATA_EXFILTRATION]: The health_check.sh script performs diagnostic checks on the local environment. It verifies the existence of necessary configuration files (.env, .codememory/config.json) and checks for the presence of required environment variables like OPENAI_API_KEY. An inline Python routine is used to test connectivity to a configured Neo4j database endpoint via a socket check; this operation is local to the network and does not transmit sensitive data to external or untrusted domains.

agentic-memory-adapter