chart — Claude Code command: decision map before capture
chart turns one oversized or fuzzy idea into a decision map. Each run settles a single decision and leaves a brief you can capture as a task.
MIT ★ 730 updated 2026-10-29
/plugin marketplace add gmickel/flow-next
/plugin install flow-next@flow-nextWhat chart does
The command is thin on purpose. All it does is pass your arguments to the flow-next-chart skill, which then picks the mode: charting, working the chart, reporting status, or re-entering from a tracker link.
One decision per invocation. Grounding against project context comes first, chart edits go through flowctl, and the skill stops at the gates that need a person. The run ends with a CHART_VERDICT= line. Arguments take a plain idea, a <chart-id> with --decision <n> or --status, or a <tracker-url>.
Because the frontmatter sets disable-model-invocation: true, the model cannot start this by itself — you do. On Claude the command file sits at .claude/commands/chart.md; on Codex it goes to ~/.codex/prompts/chart.md.
What it helps with
- split a vague idea into decisions before writing a task
- settle one decision per run instead of drowning in a backlog
- resume a chart by
<chart-id>where you stopped - check chart status without opening the tracker
- get a brief that turns into a captured task
Example prompts for the agent
/chartTake the “live in-app notifications” idea and map the decisions it hides./chart chart-3 --decision 2— show that decision and record the call./chart chart-3 --status— where did we leave off?
How to install chart
1Plugin marketplace
/plugin marketplace add gmickel/flow-next/plugin install flow-next@flow-next2Manually: put the files in place
.claude/commands/chart.mdCall it with /chart. For all projects use ~/.claude/commands/
Manually: put the files in place
~/.codex/prompts/chart.mdCall it with /prompts:chart. Codex marks this format deprecated in favour of skills
Manually: put the files in place
.github/prompts/chart.prompt.mdManually: put the files in place
.gemini/commands/chart.tomlGemini stores commands as TOML: the prompt text goes into the prompt field
Questions about chart
How do I install chart in Claude Code?
Run /plugin marketplace add gmickel/flow-next, then /plugin install flow-next@flow-next. The command file lands at .claude/commands/chart.md.
What does chart do?
It maps one large or unclear idea into decisions and settles a single decision per run. The run ends with a brief you can turn into a captured task.
Can the agent invoke chart on its own?
No. disable-model-invocation: true is set in the frontmatter, so the command is started by hand.
Source text
chart.md · MIT
--- name: chart description: Decision-map discovery for one oversized/unclear idea before capture - ground, chart a frontier, resolve one decision per invocation, brief for capture argument-hint: "[idea | <chart-id> [--decision <n>] [--status] | <tracker-url>]" disable-model-invocation: true --- # IMPORTANT: This command MUST invoke the skill `flow-next-chart` The ONLY purpose of this command is to call the `flow-next-chart` skill. You MUST use that skill now. **Arguments:** $ARGUMENTS Pass the arguments to the skill. The skill handles mode detection (chart / work / status / locator re-entry), bounded grounding, prompt-first operation inference, flowctl chart mutations, attended gates, and the terminal `CHART_VERDICT=` line.
© Gordon Mickel · MIT · gmickel/flow-next