scriptwriter-open

The user wants to open a depth-1 RRR cascade with a peer.

Parse $ARGUMENTS:

• First whitespace-separated token = peer base URL (required)
• Remaining tokens = optional basis text (defaults to "I see your work.")

Path A — MCP tools available

Call mcp__scriptwriter__open_cascade_with_peer with { peer_base_url, basis_text? }. The tool:

1. Signs the depth-1 turn locally over guild-rrr-escalate/v1 canonical bytes
2. Pushes to the peer's /rrr/turn
3. Stores the cascade locally so future escalate_cascade calls find it

Report back: the cascade ID, the depth-1 status, the peer's handle, the depth_bundle (label + tier + emoji ladder). Highlight that the ball is now in the peer's court — they must send the depth-2 turn before the chain can grow.

Path B — CLI

If MCP isn't configured:

bun packages/scriptwriter/bin/scriptwriter.ts open $ARGUMENTS

The CLI signs + pushes + prints the cascade ID and the peer's handle.

After opening

Surface the depth-1 label and the next move:

Cascade <short-id> opened with <peer-handle> at depth 1 ("they know you"). The substrate is now waiting for their depth-2 turn — you'll see it land at your /rrr/turn endpoint (if you're serving HTTP) or via the next list_cascades poll. Per wall/rrr-must-alternate, you cannot escalate from your side until they respond.

If the peer refused with rrr_cascade_already_active, there's already an open cascade between the two DIDs — point the user at list_cascades to find and continue the existing one.

If the peer refused with invalid_signature, the canonical bytes didn't match — usually a clock skew or a basis-text mismatch. Re-running typically fixes it.