Added a Conditional PR Approval Skill ht-pr-approve to hermes-talaria — Separating Multi-Model Review, Natural-Language Policy, and Approval Submission to Approve on a Pinned Commit

Lining up four model reviews still leaves the actual approve-or-not call sitting on your desk. ht-pr-approve lets an agent take that call, but only when the user supplies an explicit natural-language policy and only against a pinned commit. This post walks through why review, adjudication, and submission are kept as separate stages, why the detached worktree is not a security boundary, why 'No P0' must never silently become 'No P0/P1', and why an existing APPROVED review at the same head is treated as an idempotent no-op.

Reviewing a PR with Claude Code and Codex at Once — Automating Multi-Model Review as a Hermes Agent Skill

A code review from a single model inherits that model's blind spots wholesale. So have independent models read the same diff, and trust the findings where they overlap. Here is how that workflow became a single command.