status — Plan State Reader

SkillProductivity

Surfaces the current state of work in gflow-cli. Three variants at different levels of detail: full state (status), next task only (next), active plan identity (active). All variants run scripts/dev/active_plan.py and filter its output to the requested detail level.

Use status — Plan State Reader in Claude, ChatGPT or Ahel Desktop

Free. Sign in, add status — Plan State Reader and connect your AI. About a minute.

Also: Claude Code · Cursor · Codex

Then ask your AI: use the status skill

Details

Instructions available. Your AI can read the instructions. Execution depends on the setup they require.

Add ahel to your AI once: Claude, ChatGPT, Cursor, Claude Code or Codex. Then ask it to use this.

status — Plan State ReaderStart free

What this skill tells your AI

The instructions your AI receives, as published by ffroliva/gflow-cli in skills/status/SKILL.md and read by ahel’s review.

Three variants for different contexts. All backed by scripts/dev/active_plan.py.


Variants

status [feature] — Full state

The full picture: which plan file is active, its goal, progress (X/N tasks complete), and the next unchecked task block.

Script invocation:

# With a feature name (from $ARGUMENTS or conversation context):
uv run python scripts/dev/active_plan.py --feature <slug>

# Without a feature name (uses root PLAN.md):
uv run python scripts/dev/active_plan.py

Return: the complete script output verbatim.

When to call:

  • Starting a session: "where did we leave off?"
  • After completing a task: "what comes next?"
  • Before adding scope: "does this belong to the current plan?"

next [feature] — Next task only

The single next unchecked task block. No header noise.

Script invocation: same as status.

Return: only the content from --- Next task --- onward. Drop the Plan / Title / Goal / Progress header lines entirely.

If no --- Next task --- separator is present (root PLAN.md mode), return the full script output — the phase block is already task-level content.

If the script output contains "All steps complete", say so and suggest:

  • /gflow:changelog to review unreleased changes
  • /gflow:release if the phase is fully done

When to call:

  • "What do I do right now?" — between tasks, no orientation needed
  • Resuming mid-session after a context switch

active — Plan identity only

Just which plan is active and its goal. No task detail.

Script invocation: same as status.

Return: only the header lines — Plan path, Title, Goal, Progress count.

  • If the output contains --- Next task ---, stop before that separator.
  • If no separator is present (root PLAN.md mode — _summarise_root_plan() output), return the full output; it contains only orientation-level content with no task block.

When to call:

  • Before /gflow:predict or /gflow:scenario: confirm the proposal fits the active scope
  • Quick check: "are we in a superpowers plan or the root PLAN.md?"
  • When context is long and you need a one-line anchor without task block noise

What the script returns (reference)

  • Superpowers plan active: file path · title · goal · progress (X/N steps complete) · next unchecked task block
  • No superpowers plan: the first incomplete phase from PLAN.md (scope, sequence, definition of done)
  • All complete: "All steps complete." — prompt user toward /gflow:changelog + /gflow:release

Feature slug resolution

  1. If the caller passed a slug (e.g. shell-multi-prompt), pass it as --feature <slug>.
  2. Otherwise check conversation context for a feature name mentioned this session.
  3. If neither, run the script without --feature (falls back to root PLAN.md).

Integration

  • Claude Code: invoke via /gflow:status, /gflow:next, or /gflow:active (thin wrappers around this skill).
  • Cursor / Aider / Codex: include this file in context; call the variant by name.
  • Antigravity (agy): read this file before asking "what's the next task?".

Signals

GitHub stars
264
Forks
71
Last commit
Oct 2026
Advanced
Item type
skill
Key
status-ffroliva
Source
github.com/ffroliva/gflow-cli