The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute apm install and python scripts/validate_apm_package.py. These commands are used to deploy package artifacts and perform integrity validation as part of the intended workflow.
[PROMPT_INJECTION]: The skill processes external "APM packages" which may contain untrusted data, presenting a surface for indirect prompt injection if those packages contain instructions designed to influence the agent's behavior after deployment.
Ingestion points: Untrusted data enters the system via the packages being installed through the apm install command.
Boundary markers: The skill lacks explicit markers or instructions to treat the content of the materialized packages as untrusted or to ignore embedded instructions.
Capability inventory: The skill utilizes Bash, Read, and Glob tools to perform file system operations and execute the installation/validation logic.
Sanitization: No specific sanitization or filtering of the package contents is performed before they are materialized into sensitive directories like .agents/ or .claude/.

install-apm-package