The agent that grows with you
Python aiai-agentai-agentsanthropicchatgptclaudeclaude-codecodexhermeshermes-agentllmnous-researchopenai
"""Command descriptions survive both catalog and completion transport."""
from prompt_toolkit.completion import CompleteEvent
from prompt_toolkit.document import Document
def test_full_descriptions_survive_catalog_and_completion(monkeypatch):
from agent import skill_commands
from hermes_cli import plugins
from hermes_cli.commands_completion import SlashCommandCompleter
from tui_gateway import server
description = "Read the entire description before selecting a command. " * 8
skills = {"/proof-skill": {"name": "proof-skill", "description": description}}
monkeypatch.setattr(skill_commands, "scan_skill_commands", lambda: skills)
monkeypatch.setattr(plugins, "get_plugin_commands", lambda: {"proof-plugin": {"description": description}})
monkeypatch.setattr(server, "_load_cfg", lambda: {"quick_commands": {"proof-quick": {"description": description}}})
catalog = server._methods["commands.catalog"](1, {})["result"]
pairs = dict(catalog["pairs"])
for name in ("/proof-skill", "/proof-plugin", "/proof-quick"):
assert pairs[name] == description
completer = SlashCommandCompleter(skill_commands_provider=lambda: skills)
completions = list(completer.get_completions(Document("/proof"), CompleteEvent()))
assert len(completions) == 2
assert all(description in completion.display_meta_text for completion in completions)