The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes and analyzes existing codebase files to research product features and differentiators. This ingestion of untrusted local data creates a potential surface for indirect prompt injection if source files contain hidden adversarial instructions intended to influence the agent's code generation or implementation steps.
Ingestion points: The agent reads codebase source files and git log history as part of its research phase (SKILL.md).
Boundary markers: The instructions do not define specific delimiters or "ignore" directives for content read from the codebase.
Capability inventory: The skill has the capability to write new frontend files, update existing code, and execute development-related commands like type checkers (SKILL.md).
Sanitization: No explicit sanitization or validation of the text content extracted from the codebase is specified.
[COMMAND_EXECUTION]: The skill workflow involves executing standard development and version control commands to perform its tasks.
Evidence: The agent is instructed to use git log for product research and to run type checkers to verify code integrity before shipping (SKILL.md).
[EXTERNAL_DOWNLOADS]: The README provides instructions for installing the skill and its parent marketplace via the platform's plugin system.
Evidence: Directs the user to add the ooiyeefei/ccc marketplace and install ccc-skills@ccc. These are standard installation procedures for the vendor's toolset and do not involve untrusted third-party script execution.

landing-page-gtm