The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute shell commands such as grep, find, awk, and git. These commands are used to discover, search, and update local knowledge bases. While these operations are core to the skill's purpose, they provide a mechanism that could be misused if the agent's behavior is influenced by adversarial input.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes content from external files without sanitization or strict boundary markers.
Ingestion points: The skill reads Markdown, HTML, and Text files from paths defined in ~/.superskills/knowledge.conf.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to treat the content of the knowledge base files as data rather than instructions.
Capability inventory: The skill has access to the Bash, Write, Read, Grep, and Glob tools, which allow for file system modification and command execution.
Sanitization: The instructions do not specify any validation or filtering for the content retrieved from the repositories.

kb-advisor