Skip to main content
skill CX Codex

flow-next-deps — Claude Code skill: spec dependency graph

flow-next-deps shows spec dependency graph, execution order and blocking chains. Use it when you ask 'what's blocking what' or 'what order should specs run'.

MIT ★ 730 updated 2026-10-29

Install into Claude Code · Plugin marketplace All methods and platforms ↓
/plugin marketplace add gmickel/flow-next
/plugin install flow-next@flow-next
Open on GitHub

What flow-next-deps does

The skill builds three reports: which specs are ready, which are blocked and by what, and how to distribute them into parallel execution phases. It runs three steps: gathering all spec data via flowctl show, analyzing blocks in jq, and computing phases. If a spec appears in no phase, it's either a dependency cycle or a reference to a missing or closed spec — the skill lists those separately.

Requires jq and flowctl (bundled with the plugin). Read-only: to change dependencies, use flowctl spec add-dep and rm-dep manually.

What it helps with

  • see which specs can start right now
  • find the project's critical path
  • detect dependency cycles or references to missing specs
  • distribute specs into parallel execution phases

Example prompts for the agent

  • Show the dependency graph for all specs
  • Which specs are blocked and by what?
  • Distribute specs into execution phases

How to install flow-next-deps

1Plugin marketplace

/plugin marketplace add gmickel/flow-next
/plugin install flow-next@flow-next

This installs the whole flow-next plugin with all its skills

2Skills CLI (Vercel)

npx skills add gmickel/flow-next --skill flow-next-deps -a claude-code

3Manually: put the files in place

~/.claude/skills/flow-next-deps/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/flow-next-deps/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about flow-next-deps

How to install flow-next-deps in Claude Code?

Run /plugin marketplace add gmickel/flow-next, then /plugin install flow-next@flow-next, then npx skills add gmickel/flow-next --skill flow-next-deps -a claude-code. The skill path appears in ~/.claude/skills/flow-next-deps/SKILL.md.

What does flow-next-deps do?

Builds a spec dependency graph: shows which are ready, which are blocked, distributes them into execution phases, and finds cycles or broken references.

What does flow-next-deps require?

Requires jq (install via brew install jq) and flowctl — the script is bundled with the plugin. Without jq, the skill won't run.

Source text

SKILL.md · MIT
---
name: flow-next-deps
description: "Show spec dependency graph and execution order. Use when asking 'what's blocking what', 'execution order', 'dependency graph', 'what order should specs run', 'critical path', 'which specs can run in parallel'."
---

# Flow-Next Dependency Graph

Visualize spec dependencies, blocking chains, and execution phases.

## Preamble

flowctl is bundled with the plugin (not on PATH). Define once; subsequent blocks use `$FLOWCTL`:

```bash
FLOWCTL="${DROID_PLUGIN_ROOT:-${CLAUDE_PLUGIN_ROOT}}/scripts/flowctl"
[ -x "$FLOWCTL" ] || FLOWCTL="<plugin-root>/scripts/flowctl"   # <plugin-root> = the directory two levels above this skill's SKILL.md file (the harness gave you that file's absolute path when the skill loaded); substitute it literally
[ -x "$FLOWCTL" ] || FLOWCTL=".flow/bin/flowctl"
```

## Setup

```bash
$FLOWCTL detect --json | jq -e '.exists' >/dev/null && echo "OK: .flow/ exists" || echo "ERROR: run $FLOWCTL init"
command -v jq >/dev/null 2>&1 && echo "OK: jq installed" || echo "ERROR: brew install jq"
```

## Step 1: Gather Spec Data

Build a consolidated view of all specs with their dependencies — a single heavy per-spec loop for the whole skill. Steps 2 and 3 reuse the cached file (bash vars do not survive across tool calls, so the cache is a file at a literal agent-composed path — compose `<suffix>` once, e.g. 4 random chars, and reuse the same literal path in every later block):

```bash
# ONE gather — Steps 2 and 3 read this file; never re-run the per-spec loop
SPECS_FILE="${TMPDIR:-/tmp}/flow-deps-specs-<suffix>.json"
$FLOWCTL specs --json | jq -r '.specs[].id' | while read id; do
  $FLOWCTL show "$id" --json | jq -c '{
    id: .id,
    title: .title,
    status: .status,
    plan_review: .plan_review_status,
    deps: (.depends_on_epics // [])
  }'
done | jq -s '.' > "$SPECS_FILE"
cat "$SPECS_FILE"
```

### Done when

- `$SPECS_FILE` exists at the composed literal path and parses as a JSON array with one object per spec returned by `flowctl specs --json`.
- The per-spec `flowctl show` loop has run exactly once for this invocation. **Steps 2 and 3 read that file.** A second run of the gather loop has broken this.

## Step 2: Identify Blocking Chains

Determine which specs are ready vs blocked (pure jq, works on any shell):

```bash
# Reuse the Step 1 gather — same literal path, NO re-fetch (one heavy loop total)
SPECS_FILE="${TMPDIR:-/tmp}/flow-deps-specs-<suffix>.json"

# Compute blocking status
jq -r '
  # Build status lookup
  (map({(.id): .status}) | add // {}) as $status |

  # Check each non-done spec
  .[] | select(.status != "done") |
  .id as $id | .title as $title |

  # Find deps that are not done
  ([.deps[] | select($status[.] != "done")] | join(", ")) as $blocked_by |

  if ($blocked_by | length) == 0 then
    "READY: \($id) - \($title)"
  else
    "BLOCKED: \($id) - \($title) (by: \($blocked_by))"
  end
' "$SPECS_FILE"
```

### Done when

- Every non-`done` spec in `$SPECS_FILE` emitted exactly one `READY:` or `BLOCKED:` line, and each `BLOCKED:` line names the specific deps holding it.
- The jq ran against the cached file, not a fresh `flowctl show` sweep.

## Step 3: Compute Execution Phases

Group specs into parallel execution phases:

```bash
# Reuse the Step 1 gather — same literal path, NO re-fetch (one heavy loop total)
SPECS_FILE="${TMPDIR:-/tmp}/flow-deps-specs-<suffix>.json"

# Phase assignment algorithm (run in jq for reliability)
jq '
  # Build status lookup
  (map({(.id): .status}) | add // {}) as $status |

  # Filter to non-done specs
  [.[] | select(.status != "done")] as $open |

  # Assign phases iteratively
  reduce range(10) as $phase (
    {assigned: [], result: [], open: $open};

    .assigned as $assigned |
    .open as $remaining |

    # Find specs not yet assigned whose deps are all done or in earlier phases
    ([.open[] | select(
      ([.id] | inside($assigned) | not) and
      ((.deps // []) | all(. as $d | $status[$d] == "done" or ($assigned | index($d))))
    )] | map(.id)) as $ready |

    if ($ready | length) > 0 then
      .result += [{phase: ($phase + 1), specs: [.open[] | select(.id | IN($ready[]))]}] |
      .assigned += $ready
    else . end
  ) |
  # Emit the phases AND the residue: any open spec never assigned is UNRESOLVABLE — a
  # dependency cycle (A→B→A), a dep on a missing/closed spec, or a chain deeper than 10.
  # Dropping it silently is the one way /deps gives a WRONG answer (the graph it exists to
  # expose hides the deadlock). Surface it, with the offending deps for diagnosis.
  .assigned as $asg |
  { phases: .result,
    deadlocked: [ $open[] | select(.id as $i | ($asg | index($i)) | not)
                  | { id, status,
                      unresolved_deps: [ (.deps // [])[] | select(. as $d | ($asg | index($d)) or ($status[$d] == "done") | not) ] } ] }
' "$SPECS_FILE"
```

### Done when

- The jq result carries both `.phases` and `.deadlocked`, and every open spec appears in exactly one of them.
- **The report is rendered from `.phases` and `.deadlocked`.** Phases narrated from a reading of the spec titles have broken this.

## Output Format

Present results as:

```markdown
## Spec Dependency Graph

### Status Overview

| Spec | Title | Status | Dependencies | Blocked By |
|------|-------|--------|--------------|------------|
| **fn-1-add-auth** | Add Authentication | **READY** | - | - |
| fn-2-add-oauth | Add OAuth Login | blocked | fn-1-add-auth | fn-1-add-auth |
| fn-3-user-profile | User Profile Page | blocked | fn-1-add-auth, fn-2-add-oauth | fn-2-add-oauth |

### Execution Phases

Render from the jq result's `.phases`:

| Phase | Specs | Can Start |
|-------|-------|-----------|
| **1** | fn-1-add-auth | **NOW** |
| 2 | fn-2-add-oauth | After Phase 1 |
| 3 | fn-3-user-profile | After Phase 2 |

### ⚠️ Deadlocked / Unresolvable

**This section renders exactly when `.deadlocked` is non-empty** — and when it does, it is the
most important part of the report. Each entry is an open spec that could not be placed in
any phase: a dependency **cycle**, a dep on a **missing/closed** spec, or a chain deeper
than 10. These are invisible to `ready`/`flow --auto` (they just never become ready) — this is the
one place the graph surfaces them.

| Spec | Status | Unresolved deps | Likely cause |
|------|--------|-----------------|--------------|
| fn-7-x | blocked | fn-9-y | fn-9-y not found / closed, or a cycle fn-7↔fn-9 |

For each, state the likely cause: if two deadlocked specs list each other → **cycle** (fix
with `flowctl spec rm-dep`); if an unresolved dep isn't among the open specs → **missing or
closed dependency**. **Every entry of `.deadlocked` reaches the report.** A report that lists phases while silently dropping an open spec has broken this.

### Critical Path

fn-1-add-auth → fn-2-add-oauth → fn-3-user-profile (3 phases)
```

**This skill is read-only inspection.** Edge changes are reported as commands for the
operator to run — `flowctl spec add-dep <spec> <dep>` / `rm-dep <spec> <dep>`. A run that
executes either command itself, or edits a spec file, has broken this.

## Quick One-Liner

For a fast dependency check:

```bash
$FLOWCTL specs --json | jq -r '.specs[] | select(.status != "done") | "\(.id): \(.title) [\(.status)]"'
```

## When to Use

- "What's the execution order for specs?"
- "What's blocking progress?"
- "Show me the dependency graph"
- "What's the critical path?"
- "Which specs can run in parallel?"
- "What should I work on next?"

© Gordon Mickel · MIT · gmickel/flow-next

Similar skills

All skills for Agent setup: memory, context, orchestration →

From the same set70

The whole set →
security-scout security-scout inspects GitHub branch protection, CODEOWNERS, secret scanning, and automated dependency updates before agent tasks begin. CX Codex ★ 730repo stars flow-next-tracker-sync flow-next-tracker-sync projects a flow-next spec onto a Linear, GitHub, GitLab or Jira issue and keeps body, status and comments in sync, with the spec staying authoritative. CX Codex +1 ★ 730repo stars claude-md-scout claude-md-scout audits CLAUDE.md and AGENTS.md files on an eight-point substance scale, verifying real commands and repo specifics before agents start working. CX Codex ★ 730repo stars flow-gap-analyst The flow-gap-analyst subagent examines a short feature brief for unclear requirements, missing steps, and edge cases. Use it to surface questions before implementation begins. CX Codex ★ 730repo stars flow-next-plan-review The flow-next-plan-review skill routes Flow specs and plans to a selected reviewer, then handles fixes and a follow-up review. Use it to check a Flow document before moving ahead. CX Codex +1 ★ 730repo stars flow-next-impl-review Runs John Carmack-style implementation review via a chosen backend — claude, codex, copilot, cursor or a human. Invoked with /flow-next:impl-review. CX Codex +1 ★ 730repo stars flow-next-work flow-next-work executes Flow-Next specs and tasks step by step: manages Git branches, tracks task states in .flow/ via flowctl, runs tests, and creates clean commits. CX Codex +1 ★ 730repo stars flow-next-spec-completion-review flow-next-spec-completion-review is a skill that checks whether a spec's tasks together meet its requirements. Run it before marking work on the spec complete. CX Codex +1 ★ 730repo stars

Project sponsors

Companies that keep this analytics open