livespec-next

Ranks the next spec-side action over proposed changes and history, emitting structured JSON.

Updated Aug 15, 2026
One-click install
npx skills add https://github.com/thewoolleyman/livespec-driver-pi --skill livespec-next-thewoolleyman
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: livespec-next
Source: https://github.com/thewoolleyman/livespec-driver-pi/tree/main/skills/livespec-next
Command: npx skills add https://github.com/thewoolleyman/livespec-driver-pi --skill livespec-next-thewoolleyman

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When maintaining a livespec-governed project, it is not always obvious which spec-side operation to run next. This Skill inspects the current state of proposed changes and history, then ranks the next action (revise, propose-change, critique, prune-history, or none) so you always know what to work on. ## Core Features & Use Cases - Action Ranking: Evaluates the spec target's proposed_changes/ and history/ directories and recommends the single most appropriate next operation. - Structured JSON Output: Emits machine-readable JSON, making it usable as a primitive inside cross-repo loop drivers or automation. - Read-Only Operation: Never mutates spec state, so it is safe to run at any point in the workflow. - Use Case: After a round of critiques and revisions, ask what to do next; the Skill runs the configured next CLI and surfaces a ranked recommendation such as "revise proposal X" with supporting detail. ## Quick Start Ask the agent to run the livespec next operation and tell you what spec-side action to take next on this project.

Frequently Asked Questions about livespec-next

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I decide what spec change to work on next in livespec?▼

Run the livespec-next operation, which inspects the spec target's proposed_changes/ and history/ directories and ranks the next action: revise, propose-change, critique, prune-history, or none. The result is emitted as structured JSON.

How does the livespec next CLI get configured?▼

The governed project's .livespec.jsonc names the CLI under spec_clis.next as an argv array. If the key is absent, the Skill falls back to core's reference default: python3 <core-root>/scripts/bin/next.py.

Does livespec-next modify my specification files?▼

No, the operation is read-only. It only reads the proposed_changes/ and history/ state and emits a ranked recommendation as JSON, so it is safe to run at any point.

Why does livespec-next fail silently under pi -p non-interactive runs?▼

Pi's project-trust gate silently ignores project-local packages in non-interactive mode until a trust decision exists. The resolver's diagnostic should be surfaced verbatim rather than improvising an install.

Can livespec-next be used inside an automated loop driver?▼

Yes, its structured JSON output is designed to be composed as a primitive by a cross-repo loop driver, which can parse the ranked action and dispatch the corresponding operation skill.