The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the featbit CLI tool to perform flag management operations, such as featbit flag create, featbit flag toggle, and featbit flag set-rollout. These commands are standard for the skill's intended purpose of managing feature exposure.
[EXTERNAL_DOWNLOADS]: The skill references the installation of additional SDK-specific skills from the author's official GitHub repository (featbit/featbit-skills) using the npx skills add command. This is documented as a standard way to extend the agent's capabilities within the FeatBit ecosystem.
[PROMPT_INJECTION]: The skill ingests data from external project state (via the project-sync tool) to define rollout strategies. While this introduces a surface for indirect prompt injection, the impact is confined to the configuration of feature flags and rollout schedules within the authorized vendor environment.
Ingestion points: Project data retrieved from project-sync (specifically the get-project command output fields: goal, hypothesis, and constraints).
Boundary markers: The instructions do not define explicit delimiters or 'ignore' instructions for the ingested project data.
Capability inventory: Command execution via the featbit CLI tool for flag lifecycle management.
Sanitization: No explicit sanitization or validation of the ingested project fields is performed before they are used to generate specs or commands.

reversible-exposure-control