@longtable/cli

Researcher-facing CLI for LongTable.

LongTable keeps scholarly project state in .longtable/ and exposes generated provider skills for Codex and Claude Code. The CLI installs setup, state, checkpoint, search, panel, and diagnostic tooling. It does not replace the provider.

Install

npm install -g @longtable/cli

Primary Flow

longtable setup --provider codex
cd "<research-folder>"
codex

Then invoke:

$longtable-start

$longtable-start creates or resumes the workspace, asks open research-start questions, and stores a Research Specification when there is enough material.

After a Research Specification exists, use:

$longtable-interview

$longtable-interview is post-start. It uses option-first follow-up choices for spec revisions, checkpoint resolution, evidence boundaries, coding rules, method choices, and protected decisions. If no usable Research Specification exists, it must route to $longtable-start.

Workspace Artifacts

<project>/
  AGENTS.md
  CURRENT.md
  .longtable/
    project.json
    current-session.json
    state.json
    sessions/

• CURRENT.md: human-readable current project state
• .longtable/state.json: durable memory, questions, decisions, interview turns, evidence records, First Research Shape, and Research Specification
• QuestionRecord -> DecisionRecord: the durable checkpoint lifecycle

Provider UI is transport. LongTable supports MCP/native structured elicitation, interactive TTY selector surfaces, and numbered/plain-text fallback. Tmux is not required for LongTable core behavior.

Commands

longtable setup --provider codex
longtable doctor
longtable status --cwd "<project-path>"
longtable resume --cwd "<project-path>"
longtable roles
longtable question --prompt "<decision context>"
longtable decide --question <id> --answer <value>
longtable spec read --cwd "<project-path>"
longtable search --query "<topic>"
longtable panel --prompt "review this measurement plan" --json

longtable start remains available for scripted workspace creation with --no-interview --json, but it is not the primary research-start surface.

Panel Orchestration

Panel orchestration is for moments where disagreement matters: methods risk, measurement validity, theory fit, literature positioning, and claims that need challenge before they become project memory.

The CLI creates a provider-neutral PanelPlan and returns a planned PanelResult. When native subagents are unavailable, LongTable uses a stable sequential fallback prompt. That keeps the same research semantics available in Codex and Claude Code without making either provider's native question or agent tool the source of truth.

longtable panel --prompt "Review this measurement plan." --role editor,measurement_auditor --json
longtable panel --visibility always_visible --prompt "Keep unresolved disagreement visible." --json
longtable ask --prompt "lt debate: Review this design before I commit it." --json

Team-style requests route through panel. Explicit debate-language requests write panel debate records under .longtable/panel/; LongTable team execution is disabled for new work.

Development

npm run build --workspace @longtable/cli
npm run typecheck --workspace @longtable/cli

Codex hard-stop diagnostics

Codex Stop blocks only active LongTable hard-stop blockers: unresolved Research Specification question, scope, construct, method, evidence, or protected decision commitments. Use:

longtable codex hook-doctor --json
longtable codex status --json
longtable doctor --json

to inspect hook coverage/trust plus stopWouldBlock, activeBlockers, stale pending-question counts, and next actions. Tmux remains an optional terminal transport; LongTable state and hooks own the behavior.