This repository's own brand voice, discovered from what it already shipped.
The sister repo (knowledge-engineering) has a real skill for this: cowork/skills/plugins/brand-voice:discover-brand (SKILL.md). Its job is enterprise brand-material discovery — fan out across Notion, Slack, Google Drive, a company wiki, whatever platforms are connected to the session, triage what's found, fetch the highest-signal sources in full, and write a grounded discovery report.
This repository has none of those connectors wired in. No Notion, no Slack, no Drive. So subagentbrands.com adapts the same methodology to the material that actually exists here: the real README files, concept pages, and homepage copy already committed across every sibling worker — subagentcache.com, subagentcontracts.com, subagentidentities.com, subagentcitations.com, and the rest. Discovery here means reading this repo's own shipped pages, not fetching an enterprise platform that isn't connected.
6 guidelines, live from D1
What got recorded
Six guideline rows, each grounded in a real, cited source from this repository: a voice traits statement, a do rule, a don't rule, a terminology glossary, a good example lifted verbatim from a shipped page, and a bad example written as a deliberate hypothetical contrast — the anti-pattern this family avoids, not something any sibling site actually said.
Explore
Browse every guideline, grouped by section → · Read the full discover-brand adaptation → · Hit the JSON API →