feature

A guided workflow for planning and building software features or fixing bugs from a written specification. It records project context, decisions, plans, reviews, and audit results in project files.

In plain words
What is it for?
Use it to turn a request into acceptance criteria, research and compare solutions, plan phased implementation, review each phase, and check the finished work. It can also handle bug-fix requests.
Why use it?
It prevents work from starting without understanding the existing code and keeps each stage visible for approval. It also reduces the risk of skipping requirements or introducing changes without review.

Skill for Claude CodeCodex

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 skills/tandnguyendev/dev-workflow/feature
Any agent
npx skills add tandnguyendev/dev-workflow --skill feature
Clone the repo
git clone --depth 1 https://github.com/tandnguyendev/dev-workflow

Made for: Claude Code, Codex.

Per session 91 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 9,681 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 $0.00091 $0.09681
Opus 5 $0.00046 $0.04841
Sonnet 5 $0.00018 $0.01936
Haiku 4.5 $0.00009 $0.00968

Measured yesterday against content hash 5434d38eeca2, 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 yesterday.

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.

skills/feature/SKILL.md · 574 lines

How it starts

The opening of the file, as written. The whole thing — 574 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Feature workflow (orchestrator)

You are the ORCHESTRATOR. You delegate research, review, and coding to subagents and keep the conversation with the user — you don't do those yourself.

Golden rules

  • Source of truth is FILES, not this conversation: conventions.md and project-map.md at the repo root, and per-feature spec.md / plan.md / phase-log.md in the active feature dir .dev-workflow/features/<slug>/ (those three names always mean the copy there). Write decisions to files as you go; re-read if unsure.
  • conventions.md says HOW code is written here; project-map.md says WHAT already exists and where. Never plan a feature without knowing what the project already does — from the map, or from the survey when there is no map. An agent that doesn't know rebuilds it.
  • STOP at every checkpoint and wait for the user. NEVER skip one or advance an unapproved phase.
  • Use the feature description from arguments if given; otherwise ask first.
  • --bug in the arguments — or a request that is plainly a report of something already broken — switches Stage 0.5 to its Bug mode branch. Nothing else in the workflow changes: same tiers, same phases, same hooks, same checkpoints.

The phase-log checkboxes are MACHINE-PARSED — write them literally

phase-log.md is not just prose for humans. The hooks parse it with regexes, and the exact strings below are load-bearing. Keep every phase section in the shape templates/phase-log.md defines; a phase you write up in freeform prose is a phase the safety hooks cannot see, and they fail SILENTLY when they can't see it.

String Who writes it When What reads it
## Phase N — <title> you at scaffold all hooks — a heading that doesn't start with ## Phase or ## Final review (case-insensitive) is not a section, and its content is absorbed into the phase above
[x] coded the coder when it finishes implementing
- Evidence: <cited proof> the coder pre-fills what it RAN (test/command output); you AUGMENT with what the reviewers verified coder on return, you after reviews, before yielding the Stop evidence gate — an empty or placeholder ledger is refused
[x] code-reviewed (Final review uses full code review) you after code-reviewer returns AND its findings are resolved the Stop evidence gate — ticking either string ARMS it: from here on, ending your turn with an empty - Evidence: line is blocked
[x] security-scanned you after security-scan-fast returns; leave unticked when the phase's surface didn't warrant a scan
[x] USER APPROVED you, but ONLY after the user actually said so at the approval checkpoint status.py, and the evidence gate's "which phase is current" logic

Read the full file on GitHub · 574 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. yesterday First seen · 574 lines · 0 tokens per session scan A 5434d38eeca2

Subscribe to this mod's changes

feature is a skill published in the GitHub repository tandnguyendev/dev-workflow (2 stars, last pushed 7d ago), licensed MIT. It adds 91 tokens to every session and 9,681 once invoked, about $0.0005 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 2026-08-31.

Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

brainstorming

You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.

obra/superpowers · 37 tokens

chat-pet-sprite-creation

Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.

microsoft/vscode · 53 tokens

cpu-profile-analysis

Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…

microsoft/vscode · 71 tokens

agent-host-chat-contributions

Build and review cross-cutting agent-host chat behavior through lifecycle contributions. Use when adding turn lifecycle side effects, prompt or context injection, restored-history transformation, protocol-action observation, or when reviewing changes that add code to AgentSideEffects or AgentService.

microsoft/vscode · 56 tokens

auto-perf-optimize

Run agent-driven VS Code performance or memory investigations. Use when asked to launch Code OSS, automate a VS Code scenario, run the Chat memory smoke runner, capture renderer heap snapshots, take workflow screenshots, compare run summaries, or drive a repeatable scenario before heap-snapshot analysis.

microsoft/vscode · 62 tokens