The Agent Skills Directory

[COMMAND_EXECUTION]: Instructions specify the use of shell commands cat and gh to inspect local plugin configuration files (~/.claude/plugins/cache/) and query remote GitHub repository data for version comparisons.
[PROMPT_INJECTION]: Indirect prompt injection surface identified due to the ingestion and processing of data from an external knowledge graph.
Ingestion points: Data is retrieved from the NKS server via tools like nks_search and nks_look as defined in SKILL.md.
Boundary markers: Absent; there are no instructions to use specific delimiters or to disregard embedded directives within the graph content.
Capability inventory: The agent possesses the capability to modify graph nodes (nks_update, nks_arrow) and execute shell commands.
Sanitization: Absent; the skill does not define filtering or escaping mechanisms for the ingested data.

entry