Set modelProfile cheap only for mechanical, bounded tasks. Leave unset
(judgment/reasoning/QA) for standard tier. When in doubt, leave unset.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
- safety: drop tools section (moved to sdlc), add explicit kubectl-prod
ban, add no-tofu-direct rule, drop the merge-gate cross-reference into
a separate bullet
- sdlc: add Phase 0 product-analysis intake (CMPO Pawla as gate); add
scheduled penetration testing program (Barkley owns); standardize
authentication to Better-Auth + Google + Apple + Authentik; add
canonical tools section (moved from safety) including ghcr.io/groombook
registry standard; reorganize PR review sections to match the cross-org
pattern (named SDLC pipeline phases)
These were stale snapshots of skills owned by other orgs (better-auth,
fluxcd, greptileai, paperclipai, etc.) — Paperclip imports those
directly from their source repos at runtime. groombook/org should
contain only GroomBook-authored skills.
Mirrors the privilegedescalation/org pattern: extract company-wide
policy that was previously inlined in each agent's AGENTS.md into three
shared skills. Agents will reference these via one-line invocation
reminders in their Wake additions section.
Export all agent configs, skills, and company metadata from the
Paperclip control plane to match current GroomBook org state.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Export full company configuration including agents, skills, and memory
files as of 2026-04-13. Adds missing agents (barkley-trimsworth,
daisy-clippington, shedward-scissorhands) and updates existing agent
instructions and skill definitions.
Co-Authored-By: Paperclip <noreply@paperclip.ing>
Chris Farhood
Flea Flicker
Scrubs McBarkley