Skills¶
Skills are Claude Code slash commands bundled in the lightcone-cli plugin. They give the agent a structured, phase-by-phase workflow for the most common research operations.
If you're a researcher trying to use these, the Claude Code Workflow page in the user guide is the friendly version. This page is for maintainers.
Available skills¶
| Skill | Command | Purpose |
|---|---|---|
| lc-new | /lc-new |
Scope a research question into an astra.yaml, with optional literature extraction. |
| lc-build | /lc-build |
Plan + autonomous loop until all outputs in a universe materialize. |
| lc-verify | /lc-verify |
Read-only audit: spec validity, materialization status, decision-code alignment, result file shapes. |
| lc-migrate | /lc-migrate |
Wrap an existing codebase in ASTRA: scan, generate spec, parameterize, run. |
| lc-feedback | /lc-feedback |
File a GitHub issue against the right Lightcone repo with auto-collected context. |
How a skill is wired¶
Each skill is a claude/lightcone/skills/<name>/SKILL.md file with
YAML frontmatter:
---
name: lc-build
description: >
Build an ASTRA analysis from spec to materialized results...
allowed-tools: Read, Write, Edit, Glob, Grep, Bash(astra:*), Bash(lc:*), ...
argument-hint: "[DESCRIPTION] [--universe NAME] [--max-iterations N]"
---
The frontmatter configures Claude Code: which tools the skill may
invoke, and what the slash command's argument hint looks like. The
body is the prompt — phase definitions, rules, references to guide
files, anti-patterns. The skill bundles its own helper scripts under
scripts/ and its loop prompt template under assets/ when relevant.
Plugin layout¶
claude/lightcone/
├── skills/
│ ├── lc-new/SKILL.md
│ ├── lc-build/{SKILL.md, assets/loop-prompt.md, scripts/setup-lc-build.sh}
│ ├── lc-verify/SKILL.md
│ ├── lc-migrate/SKILL.md
│ └── lc-feedback/SKILL.md
├── agents/lc-extractor.md # subagent definition
├── guides/ # reference docs loaded by skills
├── templates/CLAUDE.md # the project CLAUDE.md template
└── scripts/*.sh # session lifecycle hooks
The plugin is force-included into the wheel via
pyproject.toml::tool.hatch.build.targets.wheel.force-include, so
lc init finds it whether you're running from source or PyPI.
Reference guides loaded by skills¶
| File | Purpose |
|---|---|
claude/lightcone/guides/astra-reference.md |
Full astra.yaml schema reference. Loaded by lc-new, lc-build, lc-migrate. |
claude/lightcone/guides/lightcone-cli-reference.md |
CLI commands, status interpretation, failure diagnosis. Loaded by build/verify skills. |
claude/lightcone/guides/ui-brand.md |
Visual formatting conventions for skill output. |
claude/lightcone/agents/lc-extractor.md |
Literature extraction subagent invoked by /lc-new. |
Authoring a new skill¶
See Authoring Skills.