The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a bash block to execute the date command for session timestamping. While the command itself is benign, it represents the execution of system-level utilities.
[COMMAND_EXECUTION]: The skill includes an ARGUMENTS: $ARGUMENTS placeholder at the end of the file. If the host platform interpolates user-provided strings into this placeholder within a shell context without proper escaping, it could lead to command injection (e.g., if a user provides an argument like ; mal_command).
[PROMPT_INJECTION]: The skill implements an '/philosophy check' command that instructs the agent to review 'current work' against the Oracle principles. This creates a surface for indirect prompt injection, as malicious instructions embedded within the 'current work' could attempt to override the agent's behavior during the review process.
Ingestion points: User-provided arguments via $ARGUMENTS and external content analyzed during the alignment check.
Boundary markers: The skill lacks explicit delimiters or instructions to ignore embedded commands when processing these inputs.
Capability inventory: The skill references tools such as arra_trace(), arra_supersede(), and AskUserQuestion().
Sanitization: There is no evidence of sanitization, filtering, or validation for the interpolated data.

philosophy