Skip to content
Agentmart
Submit a tool

status

Surfaces the current state of work in gflow-cli.

Skill by ffroliva

IndexedMIT★ 332
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/ffroliva-status

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

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 cases: productivity

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/status

What's inside

Skill name
status
Description
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.
License
MIT
Instructions
91 lines in SKILL.md · view on GitHub

Agent Skills spec issues

  • files: referenced but missing: scripts/dev/active_plan.py

Files (1)

  • SKILL.md

SKILL.md (first lines)

# `status` — Plan State Reader

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:**
```bash
# 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?".

From ffroliva/gflow-cli, MIT.

More from ffroliva/gflow-cli

Signals

Stars
332
Forks
95
Open issues
30
Contributors
22
Last commit (this folder)
Jul 20, 2026 (3 months ago)
Last release
Oct 24, 2026 (4 days ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from ffroliva

Cite this page

status skill. Agent Mart. https://agentmart.dev/skills/ffroliva-status (last updated 2026-10-24)