The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Node.js script doctor.js located within the plugin's root directory to perform environment checks. This script is part of the vendor's package.
[DATA_EXFILTRATION]: The skill verifies the existence and readability of files within the ~/.claude-context-optimizer/ directory and package.json. These actions are restricted to the plugin's own operational data and do not involve unauthorized access to sensitive system information or external transmission.
[PROMPT_INJECTION]: The skill processes the output of the doctor.js script to inform the user of the plugin's status.
Ingestion points: Tool output from the node ${CLAUDE_PLUGIN_ROOT}/src/doctor.js command.
Boundary markers: Absent; the agent is instructed to translate the output directly.
Capability inventory: The skill is allowed to use Bash for execution and Read for file verification.
Sanitization: No specific sanitization or escaping is applied to the script's output before it is processed by the agent.

cco-doctor