release-readiness

# /release-readiness — Release Decision Composer You compose the sprint's release evidence into a single decision. You do not run any of the upstream skills; you read the artifacts they already saved and emit a status that gates `/ship`. The conductor's `phase_graph` puts you between the upstream phases and `/ship`, so this skill is the last thing that runs before delivery. This is the `release-readiness` skill from the compliance-release stack. PR 2 of the Custom Stack Examples v1 round wires the real composer logic; this PR (PR 1) ships the skill structure so the static contract validates. ## Process ### 0. Resolve paths (host-agnostic) ```bash NANOSTACK_ROOT="${NANOSTACK_ROOT:-$HOME/.claude/skills/nanostack}" SKILL_DIR="${SKILL_DIR:-$HOME/.claude/skills/release-readiness}" ``` ### 1. Resolve upstream evidence ```bash NANOSTACK_ROOT="${NANOSTACK_ROOT:-$HOME/.claude/skills/nanostack}" "$NANOSTACK_ROOT/bin/resolve.sh" release-readiness ``` The resolver returns `phase_kind: "custom"` and `upstream_artifacts` with five keys: `review`, `qa`, `security`, `license-audit`, `privacy-check`. Each value is either a path to the artifact JSON or `null` if no artifact exists for that upstream. ### 2. Compose the decision ```bash SKILL_DIR="${SKILL_DIR:-$HOME/.claude/skills/release-readiness}" "$SKILL_DIR/bin/summarize.sh" ``` The helper reads each upstream artifact through `bin/find-artifact.sh --verify` and maps each to a `check` entry. Per-check status: - **`MISSING`** for ...