The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local bash scripts using paths resolved at runtime from the plugin cache directory (~/.claude/plugins/cache/insight-wave/cogni-portfolio/). It also includes instructions for a registration command that takes user-provided project paths as arguments, which could lead to command injection if the input is not handled securely by the script.
[PROMPT_INJECTION]: The skill ingests untrusted project data from external files (e.g., portfolio.json) to influence agent behavior, including the selection of communication language and the prioritization of workflow recommendations.
Ingestion points: Project configuration and metadata files such as portfolio.json, taxonomy files, and context/context-index.json.
Boundary markers: No delimiters or explicit warnings are used to prevent the agent from processing instructions that may be embedded within these external files.
Capability inventory: The agent has access to shell execution (bash), file modification (Write, Edit), and the ability to dispatch secondary skills.
Sanitization: The instructions do not define any sanitization or validation logic for the data ingested from project files.

portfolio-resume