The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill provides tools to read and reveal plaintext secrets from the macOS Keychain using the /usr/bin/security utility. While this is the intended functionality, it involves handling and exposing sensitive local credentials.
[COMMAND_EXECUTION]: The bundled shell scripts execute the macOS system utility /usr/bin/security to perform credential management tasks such as querying, creating, and deleting items.
[INDIRECT_PROMPT_INJECTION]: The skill processes external inputs as arguments for shell commands, creating a surface for indirect injection if the processed data contains malicious content.
Ingestion points: Parameters such as --service, --account, and --secret passed to the scripts in the scripts/ directory.
Boundary markers: None. The prompt instructions do not provide delimiters or specific instructions for the agent to treat these values as untrusted data.
Capability inventory: Execution of /usr/bin/security to read, write, and delete sensitive data in the macOS Keychain.
Sanitization: The shell scripts use array expansion for command execution to prevent simple argument injection. Output intended for the agent is escaped using a custom json_escape function.
[EXTERNAL_DOWNLOADS]: The skill's documentation identifies its source as the author's repository on GitHub (github.com/flc1125/skills).

mac-keychain