normalize-and-group-imported-agent-packlisted

# Normalize and Group an Imported Agent Pack ## What This Skill Does Ingests a batch of externally-authored subagent `.md` files (typically "agent-studio" / B2B packs handed over in groups of 8–12) and folds them into this repo's `agents/` roster + `skills/<domain>/<domain>-team/` ecosystem. This is **agent/team plumbing**, not skill authoring — use `skill-builder` to write a single new skill from an interview, and `library-audit` to review skills that already exist. The pipeline recurred identically across multiple packs, so the operations are codified here rather than re-derived each time. ## Context to read first - `agents/` — the current roster you are merging against (canonical names + descriptions). - An existing team SKILL as the structural template, e.g. `skills/api/api-team/SKILL.md` — copy its section shape exactly: Overview table → Adaptive agent selection → When to Use → Phases 1–5 → Model tiering → Integration (recon-swarm + keyword hook) → Related. - `skills/README.md` — the catalogue you must update (source of truth for layout). - `scripts/install.sh` — the deploy step run at the end. ## Steps ### 1. Normalize each incoming file Strip studio/B2B framing down to a clean agent definition: - Reduce frontmatter to `name`, `description`, `model` (drop everything else). - Delete framing walls: legal disclaimers, "millions of users", "6-week sprint", marketing voice, literal `\n` escape garbage from bad exports. - Rewrite `description` in house style —