0.1.6

release 0.1.6

• feat: legal, research and clinical departments
• feat: one page for skills and connectors, filtered the way people think
• refactor: one list of what this can do, not two
• docs: say why the bundled model list is empty, rather than guessing
• test: mark the remote-model suite as covering an empty catalogue
• test: stop the harness retrying against a mock with nothing left to say
• test: mark the personality suite as covering something nothing ships
• test: follow two compaction messages that were rewritten
• test: bring the truncation suite back to what the code actually does
• fix: judge a compaction on the conversation, not on the whole request