The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references the global installation of the @creatibi/cbi-cli package from the NPM registry. This is a legitimate external dependency required for the tool's operation.
[COMMAND_EXECUTION]: The skill utilizes cbi command-line utilities to perform configuration and authentication tasks. These shell operations are within the scope of the tool's intended use.
[CREDENTIALS_UNSAFE]: The skill manages application credentials and OAuth tokens stored in ~/.cbi/config.json. It provides instructions for credential initialization and viewing, using placeholders in documentation to prevent exposure. This is standard practice for CLI authentication.
[PROMPT_INJECTION]: The skill defines triggers that respond to specific error strings such as 'Permission denied' or 'Token expired'. This creates an indirect prompt injection surface where the output of other tools can influence the agent's behavior to trigger this skill.
Ingestion points: Command outputs or error messages processed by the agent that match the defined trigger phrases in SKILL.md.
Boundary markers: No specific boundary markers are used in the instruction set to isolate external content.
Capability inventory: Shell command execution through the cbi CLI tool as defined in SKILL.md.
Sanitization: No sanitization is performed on the triggering strings as they are used for pattern matching.

cbi-shared