feature

A command for managing feature development from a written specification through deployment, using separate workflow stages and quality checks.

In plain words
What is it for?
Use it to start or continue feature work, manage epics and sprints, track workflow state, and run development phases through deployment.
Why use it?
It gives complex feature work a saved, resumable process with visible progress instead of relying on one uninterrupted session.

Command for Claude Code

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/marcusgoll/spec-flow/feature
Clone the repo
git clone --depth 1 https://github.com/marcusgoll/Spec-Flow

Made for: Claude Code.

Per session 13 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 7,718 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00013 $0.07718
Opus 5 $0.00006 $0.03859
Sonnet 5 $0.00003 $0.01544
Haiku 4.5 $0.00001 $0.00772

Measured 2d ago against content hash 27ac4b54911a, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

feature scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/commands/core/feature.md · 925 lines

How it starts

The opening of the file, as written. The whole thing — 925 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Command: /feature [feature description | slug | continue | next | epic:<name> | epic:<name>:sprint:<num> | sprint:<num>]

CRITICAL ARCHITECTURE (v5.0 - Domain Memory v2):

This orchestrator is ultra-lightweight. You MUST:

  1. Read state from disk (state.yaml, interaction-state.yaml, domain-memory.yaml)
  2. Spawn isolated phase agents via Task tool - NEVER execute phases inline
  3. Handle user Q&A when agents return questions
  4. Update state.yaml after each phase
  5. NEVER carry implementation details in your context

Benefits: Unlimited feature complexity, observable progress, resumable at any point, each phase gets fresh context.

Git status: !git status --short

Current branch: !git branch --show-current

Recent features: !ls -t specs/ 2>/dev/null | head -5

Active workflow state (if any): !find specs -name "state.yaml" -exec grep -l "status: in_progress\|status: failed" {} \; 2>/dev/null | head -3

Deployment model detection: !git branch -r | grep -q "staging" && echo "staging-prod" || (git remote -v | grep -q "origin" && echo "direct-prod" || echo "local-only")

Worktree context: !bash .spec-flow/scripts/bash/worktree-context.sh info 2>/dev/null || echo '{"is_worktree": false}'

Studio context (multi-agent isolation): !bash .spec-flow/scripts/bash/worktree-context.sh studio-detect 2>/dev/null || echo ""

Worktree preference: !bash .spec-flow/scripts/utils/load-preferences.sh --key "worktrees.auto_create" --default "true" 2>/dev/null || echo "true"

Planning depth preference: !bash .spec-flow/scripts/utils/load-preferences.sh --key "planning.auto_deep_mode" --default "false" 2>/dev/null || echo "false"

Auto-ship preference: !bash .spec-flow/scripts/utils/load-preferences.sh --key "deployment.auto_ship" --default "false" 2>/dev/null || echo "false"

Auto-merge preference: !bash .spec-flow/scripts/utils/load-preferences.sh --key "deployment.auto_merge" --default "false" 2>/dev/null || echo "false"

Auto-finalize preference: !bash .spec-flow/scripts/utils/load-preferences.sh --key "deployment.auto_finalize" --default "true" 2>/dev/null || echo "true"

<planning_depth>

Planning Depth Mode (--deep / --auto)

Flags in $ARGUMENTS:

  • --deep → Force ultrathink/craftsman planning for this feature
  • --auto → Use preferences to determine depth (respects auto_deep_mode), AND continue through ship→finalize without stopping
  • Neither → Interactive mode, respects preference triggers

Pass flag to /plan phase: When spawning the plan-phase-agent, include the planning mode:

  • If --deep in arguments: Pass --deep to /plan
  • If --auto in arguments: Determine from preferences and pass appropriate flag
  • Store planning_mode in state.yaml for reference

State tracking:

# In state.yaml
planning:
  mode: deep  # or standard
  ultrathink_enabled: true
  craftsman_decision_generated: true

</planning_depth>

<auto_mode>

Read the full file on GitHub · 925 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 925 lines · 13 tokens per session scan A 27ac4b54911a

Subscribe to this mod's changes

feature is a command published in the GitHub repository marcusgoll/Spec-Flow (91 stars, last pushed 4mo ago), licensed MIT. It adds 13 tokens to every session and 7,718 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.