The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to run node .gitnexus/run.cjs analyze in the terminal when the index is stale. This involves executing a script file located within the project's hidden directory (.gitnexus). Since the contents of this script are not provided in the skill definition and could be part of an untrusted repository, this represents an unverified execution of local code.
[PROMPT_INJECTION]: The skill facilitates indirect prompt injection by ingesting and processing content from external repositories.
Ingestion points: Code search results and execution flow data are retrieved through query, context, and cypher tools, as well as file reads via the gitnexus:// protocol.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to treat the retrieved code content as untrusted data or to ignore instructions embedded within it.
Capability inventory: The skill allows the agent to read files, perform graph database queries, and execute terminal commands, which could be exploited if malicious code influences the agent's logic.
Sanitization: No sanitization, escaping, or validation of the repository content is performed before it is integrated into the agent's context.

gitnexus-debugging