The Agent Skills Directory

[COMMAND_EXECUTION]: The gno completion install command modifies shell initialization files (e.g., .bashrc, .zshrc) to enable tab completion. This is a common persistence mechanism for CLI tools to maintain their environment across sessions.
[COMMAND_EXECUTION]: The gno mcp install and gno skill install commands modify AI assistant configuration files (e.g., claude_desktop_config.json) to register the gno command. This represents a modification of the agent's operating environment.
[COMMAND_EXECUTION]: The gno serve command binds to 0.0.0.0 by default, which can expose the document indexing service and web UI to other devices on the network.
[EXTERNAL_DOWNLOADS]: The gno models pull command downloads AI models from remote infrastructure to facilitate local embedding and AI answering capabilities.
[EXTERNAL_DOWNLOADS]: The gno index --git-pull and gno update --git-pull flags enable the tool to synchronize with remote Git repositories, which involves fetching and potentially executing content updates from external sources.
[PROMPT_INJECTION]: The skill provides an interface for Indirect Prompt Injection through its document indexing and Q&A workflows.
Ingestion points: Untrusted data enters the context through files processed by gno index and gno collection add (referenced in SKILL.md and cli-reference.md).
Boundary markers: There are no documented delimiters or instructions to the AI agent to ignore embedded commands within retrieved document snippets.
Capability inventory: The skill uses Bash(gno:*) and Read tools, providing a surface for following instructions extracted from documents.
Sanitization: The documentation does not describe any validation or sanitization of the content indexed from local files.

gno