Agent skill

Feature Dev

by sd0xdev in sd0xdev/sd0x-harness

Feature development workflow. An agent skill from sd0xdev/sd0x-harness.

MITAuto-check: notesDevelopment

Install Feature Dev

skills CLI
$ npx skills add sd0xdev/sd0x-harness --skill feature-dev -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install sd0xdev/sd0x-harness feature-dev --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/sd0xdev/sd0x-harness.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/feature-dev .claude/skills/feature-dev && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
feature-dev
GitHub stars
192
Token cost
~2.1k tokens
SKILL.md length
746 words
Files
1
Skills in repo
91
Repo updated
First seen
Licence
MIT

At a glance

Feature development workflow. An agent skill from sd0xdev/sd0x-harness.

  • Works in 3 steps: Run existing tests → Test adequacy review (mandatory for code… → Code review (auto-loop)
  • : implementing features
  • SKILL.md covers Trigger, When NOT to Use, Prohibited Actions and Workflow, plus 10 more sections
  • Calls git and node

What it does

Feature Dev is an agent skill from sd0xdev/sd0x-harness. Feature development workflow. Use when: implementing features, writing code, running dev loop. Not for: understanding code (use code-explore), reviewing code (use codex-code-review). Output: implemented feature + tests + review gate.

Its SKILL.md is about 2.1k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Development, covering Code review and Human-in-the-loop approvals. The repository describes itself as: The harness layer for Claude Code — a reference implementation of harness engineering with hook-enforced dual review, state-machine gates that survive context compaction, and… The licence is MIT.

When your agent uses it

  • : implementing features
  • Running dev loop

Example prompts

  • “/feature-dev”

Requirements

  • Pre-approved tools (allowed-tools): Read, Grep, Glob, Edit, Write, Bash, Skill, AskUserQuestion

Workflow steps

3 steps, taken from the step headings in SKILL.md.

  1. Run existing tests
  2. Test adequacy review (mandatory for code changes)
  3. Code review (auto-loop)

What it can do on your machine

Read from SKILL.md and the folder at commit c9a2036. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves these tools, so the agent can use them without asking each time:

    • Read
    • Grep
    • Glob
    • Edit
    • Write
    • Bash
    • Skill
    • AskUserQuestion

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • git
    • node

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use git, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Feature Dev loads about 2.1k tokens when it runs. Until then it costs about 61 tokens; SKILL.md has 746 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~61
When it runs · the whole SKILL.md, loaded when a task matches
~2.1k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check: notes

The automated check noted patterns worth knowing about, such as sudo or a known installer.

  • NotePre-approves every shell command (allowed-tools: Bash)SKILL.md
    allowed-tools: Read, Grep, Glob, Edit, Write, Bash, Skill, AskUserQuestion

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from sd0xdev/sd0x-harness at commit c9a2036, republished under its MIT licence (© sd0xdev). 746 words, ~2,110 tokens.

Download SKILL.mdSave it as .claude/skills/feature-dev/SKILL.md (or your agent's skills folder).
name
feature-dev
description
Feature development workflow. Use when: implementing features, writing code, running dev loop. Not for: understanding code (use code-explore), reviewing code (use codex-code-review). Output: implemented feature + tests + review gate.
allowed-tools
Read, Grep, Glob, Edit, Write, Bash, Skill, AskUserQuestion

Feature Development Skill

Trigger

  • Keywords: develop feature, implement, write code, verify, precommit, refactor, simplify

When NOT to Use

  • Just want to understand code (use Explore)
  • Review code only (use codex-code-review)
  • Review documents only (use doc-review)
  • Pure test-only tasks without feature changes (use /codex-test-review directly)

Prohibited Actions

❌ git add | git commit | git push — per @rules/git-workflow.md

This skill implements features but does not commit. /precommit is a quality gate only. To commit, offer the menu per rules/git-workflow.md § Proactive Offer — a commit option on any real branch, a push option only where review-state.js offer allows one — and invoke /smart-commit --execute on selection; never print the command for the user to copy.

budget:token_budget200000</budget:token_budget>

Workflow

Requirements → Design → Implement → Test + Review → Precommit Gate → Doc Sync
                │          │            │                  │               │
                ▼          ▼            ▼                  ▼               ▼
           /codex-     /codex-    /verify              /precommit  /update-docs
           architect   implement  /codex-test-review   (or /precommit)  /create-request --update
                                  /codex-review-fast

Load the intent first: identify the feature this work belongs to (from the task, the spec/requirements being followed, or the paths being changed) and read docs/features/<key>/intent-<key>.md if it exists — it overrides your default approach. Work that contradicts one of its INV-* invariants or Non-goals stops and asks the user (cite the line; amending intent is their re-decision). No identifiable feature → nothing to load; proceed.

Design before code: for a non-trivial change, briefly consider who owns each responsibility, the simplest shape that fits the existing code, and why — and say what you chose when the choice is non-obvious. Principles (clear names, small cohesive functions, dependency direction, composition where it reduces coupling) are questions, not quotas: never add an abstraction to demonstrate design. /codex-architect is for genuinely hard trade-offs (cross-module boundaries, new public APIs, durable abstractions), not every feature.

agentctl (optional)

At the start, run node "${CLAUDE_PLUGIN_ROOT}/skills/agentctl-setup/scripts/agentctl-setup.js" status once. Only when state is installed/enabled, offer to draft a task proposal from the ticket per @skills/agentctl-setup/references/workflow-integration.md — the user accepts it with /agentctl accept; any other state, say nothing about agentctl and continue. It is never a gate: every review and precommit obligation in this skill is unchanged, and its evidence is context beside a verdict, never the verdict.

Commands

PhaseCommandDescription
Design/codex-architectGet architecture advice
Implement/codex-implementCodex writes code
Test: Run/verifyRun tests (lint → typecheck → unit → integration)
Test: Review/codex-test-reviewMandatory — review test sufficiency (5 dimensions)
Test: Generate/codex-test-genGenerate unit tests for gaps
Test: Integration/post-dev-testWrite missing integration/e2e tests
Review/codex-review-fastCode review (auto-loop)
Precommit/precommitlint + build + test (auto-loop canonical path)
Doc Sync/update-docsSync docs with code
Doc Sync/create-request --updateUpdate request progress
Refactor/simplifyFinal refactoring

Test + Review Phase (Detail)

This is the core of feature-dev — ensuring sufficient test coverage before code review.

Step 1: Run existing tests
/verify → all tests pass?
  Yes → Step 2
  No → fix failures → re-run /verify
Step 2: Test adequacy review (mandatory for code changes)
/codex-test-review → ✅ Tests sufficient?
  Yes → Step 3
  No → close gaps (Step 2a) → /codex-test-review --continue
Step 2a: Gap closure
Gap TypeRemediation Command
Unit test missing/insufficient/codex-test-gen → write tests → /verify
Integration/E2E missing/post-dev-test → write tests → /verify
Step 3: Code review (auto-loop)
/codex-review-fast → ✅ Ready?
  Yes → Precommit Gate
  No → fix issues → re-run /codex-review-fast (auto-loop)
Show full SKILL.md (304 more words)Show less
Freshness rule

If code changes after the latest ✅ Tests sufficient gate (e.g., fixes from code review), rerun /verify then /codex-test-review --continue before proceeding to precommit gate.

Testing Requirements

Follow @rules/testing.md for conventions (AAA, naming, evidence model). Follow @rules/testing-project.md for project-specific overrides (directories, runner, adequacy mode).

Change TypeTest Requirements
New Service/ProviderMust have corresponding unit test
Modify existing logicExisting tests pass + new logic tested
Bug fixMust add regression test
New API endpointIntegration test required
Cross-service changeE2E test required

Test File Mapping

Use project convention from @rules/testing-project.md. If no override is defined, follow ecosystem defaults:

Source PatternTest Pattern
src/<module>/test/unit/<module>/ or test/<module>/
scripts/<name>.shtest/scripts/<name>.test.js
skills/<name>/SKILL.mdtest/skills/<name>.test.js

Output

  • Implemented feature code + tests
  • Test adequacy gate: ✅ Tests sufficient
  • Review gate: ✅ Ready
  • Precommit results: ✅ All Pass

Verification Checklist

  • All tests pass (/verify)
  • Test adequacy reviewed (/codex-test-review)
  • Code review passed (/codex-review-fast ✅ Ready)
  • Precommit passed (/precommit ✅ All Pass)
  • No git add/commit/push executed

Doc Sync (after precommit Pass)

⚠️ Auto-triggered by @rules/auto-loop.md — behavior-layer rule, not hook-enforced.

Only when change maps to a feature under docs/features/. Target detection uses 3-level fallback — see /update-docs for algorithm details.

precommit Pass
  → Locate feature docs (see /update-docs 3-level fallback)
  → /update-docs docs/features/<feature>/2-tech-spec.md      (current-authority doc)
  → /create-request --update docs/features/<feature>/requests/<date>-<title>.md   (record: status + outcome only)
  → /codex-review-doc            (ONE dispatch for every doc touched above)
  → Safety valve: new code diff? → back to review loop (see /update-docs)

Sync the current-authority doc; append to the record. The tech spec states what is true now, so code landing makes it stale and it is rewritten. A request ticket, review log or ADR states what was decided or done at a point in time — it is updated with status and outcome, never rewritten to match today's code, because rewriting it destroys the record. scripts/lib/doc-metadata.js decides which a file is; /codex-review-doc reviews each under the profile that classification earns.

One /codex-review-doc for all of them, not one per file: the changed docs are a single review plan (skills/doc-review/SKILL.md § Workflow). Per-file dispatch is what multiplied a three-file doc sync into three whole-document reviews.

Review Loop

MUST re-review after fix until PASS (per @rules/auto-loop.md)

Review → Issues found → Fix → Re-review → ... → ✅ Pass → Next step

Examples

Input: Implement a fee calculation method
Action: /codex-architect → /codex-implement → /verify → /codex-test-review → /codex-review-fast → /precommit
Input: This code needs refactoring
Action: /simplify → /verify → /codex-test-review → /codex-review-fast → /precommit
Input: Feature dev, continue (resuming work)
Action: Check git status → identify remaining tasks → continue from current phase

© sd0xdev, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in skills/feature-dev of sd0xdev/sd0x-harness.

Open the folder on GitHubat commit c9a2036

Compare with similar skills

Feature Dev next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Feature Dev compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Feature Dev this skillsd0xdev/sd0x-harness192—~2.1kAutomated safety check: NotesMIT
Code ReviewerCaoMeiYouRen/caomei-auth220—~1.5kAutomated safety check: PassMIT
Plannotator Referencebacknotprop/plannotator9.2k—~5.7kAutomated safety check: WarnApache-2.0
Code ReviewerMageByte-Zero/spec-superflow8391 repos~1.5kAutomated safety check: PassMIT
PR Babysitteropeninterpreter/openinterpreter69k3 repos~4.2kAutomated safety check: PassApache-2.0
Code Review ChecklistshareAI-lab/learn-claude-code78k5 repos~1.1kAutomated safety check: PassMIT

Similar skills

  • Code Reviewer

    CaoMeiYouRen/caomei-auth

    审查当前 git 变更、PR、提交范围、技能定义文件、架构调整或安全敏感代码时使用。输出结构化 Review Gate 结论(Pass/Reject)、问题分级(blocker/warning/suggest)、最低验证矩阵、证据链与复查基线。覆盖正确性、安全、架构、SOLID、可删除代码、性能、异常处理与测试风险;默认只输出 review,不直接修改代码。用户提到 review、code…

    220 GitHub stars~1.5k tokensUpdated 6 days ago
    DevelopmentAuto-check passed
  • Plannotator Reference

    backnotprop/plannotator

    Reference for picking the right Plannotator tool or command for plan review, code review, annotating files and URLs, archived plan decisions and Guided Reviews.

    9.2k GitHub stars~5.7k tokensUpdated yesterday
    DevelopmentAuto-check: warnings
  • Code Reviewer

    MageByte-Zero/spec-superflow

    Review completed implementation batches for spec compliance and code quality.

    839 GitHub starsUsed in 1 repo~1.5k tokens
    DevelopmentAuto-check passed
  • PR Babysitter

    openinterpreter/openinterpreter

    Watches an open GitHub pull request until it merges, handling review comments, diagnosing CI failures and retrying flaky checks along the way.

    69k GitHub starsUsed in 3 repos~4.2k tokens
    DevelopmentAuto-check passed
  • Code Review Checklist

    shareAI-lab/learn-claude-code

    Reviews code against a five-part checklist covering security, correctness, performance, maintainability and testing, and reports findings in a fixed format.

    78k GitHub starsUsed in 5 repos~1.1k tokens
    DevelopmentAuto-check passed
  • Backend Code Review

    langflow-ai/langflow

    Review backend code for quality, security, maintainability, and best practices based on established checklist rules.

    156k GitHub stars~3.5k tokensUpdated today
    DevelopmentAuto-check: notes

More from sd0xdev/sd0x-harness

All 91 skills in this repo
  • Adr

    sd0xdev/sd0x-harness

    Write an Architecture Decision Record (ADR) for a feature — Context / Decision / Status / Consequences / Alternatives, filed as docs/features/<feature/adr-<NNN-<title.md with a 3-digit zero-padded…

    192 GitHub stars~4.8k tokensUpdated yesterday
    Auto-check passed
  • Load PR Review

    sd0xdev/sd0x-harness

    Load GitHub PR review comments into AI session — analyze, triage, plan.

    192 GitHub stars~4.4k tokensUpdated yesterday
    Auto-check passed
  • Next Step

    sd0xdev/sd0x-harness

    Change-aware next step advisor. An agent skill from sd0xdev/sd0x-harness.

    192 GitHub stars~1.6k tokensUpdated yesterday
    Auto-check passed
  • Obsidian CLI

    sd0xdev/sd0x-harness

    Obsidian vault integration via official CLI. An agent skill from sd0xdev/sd0x-harness.

    192 GitHub stars~1.1k tokensUpdated yesterday
    Auto-check passed
  • Orchestrate

    sd0xdev/sd0x-harness

    Agent-driven workflow orchestration (v1 report-only). An agent skill from sd0xdev/sd0x-harness.

    192 GitHub stars~2.5k tokensUpdated yesterday
    Auto-check passed
  • PR Comment

    sd0xdev/sd0x-harness

    Post friendly review comments to a GitHub PR — prepare locally, preview, then submit as atomic review.

    192 GitHub stars~1.5k tokensUpdated yesterday
    Auto-check passed

Categories

Questions about Feature Dev

What does Feature Dev do?

Feature development workflow. An agent skill from sd0xdev/sd0x-harness. Feature Dev is an agent skill from sd0xdev/sd0x-harness. Feature development workflow.

When should I use Feature Dev?

Feature Dev fits situations like: : implementing features; running dev loop.

How do I install Feature Dev in Claude Code?

Run `npx skills add sd0xdev/sd0x-harness --skill feature-dev -a claude-code`. Or copy the skill folder (skills/feature-dev in sd0xdev/sd0x-harness) into .claude/skills/feature-dev in your project. Claude Code loads it when a task matches its description.

How do I install Feature Dev in Codex?

Run `npx skills add sd0xdev/sd0x-harness --skill feature-dev -a codex`. Or copy the skill folder (skills/feature-dev in sd0xdev/sd0x-harness) into .agents/skills/feature-dev in your project. Codex loads it when a task matches its description.

Can I use Feature Dev in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add sd0xdev/sd0x-harness --skill feature-dev -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/feature-dev, .gemini/skills/feature-dev, .github/skills/feature-dev and .opencode/skills/feature-dev in your project.

What does Feature Dev need to run?

Going by SKILL.md and its folder, Feature Dev needs the command-line tools its instructions call (git and node). Its frontmatter pre-approves these tools: Read, Grep, Glob, Edit, Write, Bash, Skill, AskUserQuestion.

Does Feature Dev access the network?

SKILL.md contains no URLs. Its commands use git, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Feature Dev safe to install?

Our automated static check of SKILL.md found notes only (pre-approves every shell command (allowed-tools: bash)), nothing it rates as a warning. It is not a guarantee. Review the folder before installing.

What licence does Feature Dev use?

Feature Dev is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Feature Dev use?

About 2.1k tokens (SKILL.md is roughly 8.4k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Feature Dev?

Skills that share tags, products or a category with Feature Dev: Code Reviewer (CaoMeiYouRen/caomei-auth, 220 stars), Plannotator Reference (backnotprop/plannotator, 9.2k stars), Code Reviewer (MageByte-Zero/spec-superflow, 839 stars) and PR Babysitter (openinterpreter/openinterpreter, 69k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Feature Dev?

sd0xdev (a GitHub user) maintains it in sd0xdev/sd0x-harness, which has 192 GitHub stars. The repository holds 91 skills in this directory. The repository was last updated on October 6, 2026.

Source: sd0xdev/sd0x-harness on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.