Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/zxpmail/ReqForgenpx agentmods add skills/zxpmail/reqforge/dev-builderWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/zxpmail/reqforge/dev-builder)<a href="https://agentmods.dev/skills/zxpmail/reqforge/dev-builder"><img src="https://agentmods.dev/badge/skills/zxpmail/reqforge/dev-builder/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/zxpmail/reqforge/dev-builder"><img src="https://agentmods.dev/badge/skills/zxpmail/reqforge/dev-builder.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 3 findings, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Prompt Injection · line 1 Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.Fix: Audit all comments and invisible characters. Remove any instructions that direct the agent to perform unauthorized actions. Use plain, reviewable content.
- high Prompt Injection · line 57 Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.Fix: Audit all comments and invisible characters. Remove any instructions that direct the agent to perform unauthorized actions. Use plain, reviewable content.
- medium Excessive Agency · line 136 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00000 | $0.03428 |
| Opus 5.5 | $0.00000 | $0.01371 |
| Sonnet 5.5 | $0.00000 | $0.00686 |
| Haiku 4.5 | $0.00000 | $0.00343 |
Grade A, and why
dev-builder 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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 205 lines — stays where its author put it; the contents beside it link to each section on GitHub.
name: dev-builder description: Used when DEV-PLAN.md is ready and the user says to start coding or continue developing the next Phase. Sets up the skeleton for new projects, implements features by Phase for existing projects. version: 1.2.0 updated: 2026-06-27 requires: []
[Task] Initialization Mode (0-to-1): No code + has DEV-PLAN.md -> set up project skeleton according to tech stack, install dependencies, configure development environment, complete Phase 1. Phase 1 的核心不是"写功能",是"放骨架": - 先写领域模型、核心类型定义、数据流接口 - 再写 Validator 模式、错误处理公约、中间件链 - 这些就是项目独有的"天理"——牛的骨节结构 - 后续所有 Phase 的代码都是从这些骨架生长出来的 - 所以 Phase 1 宁可慢一点把骨架放稳,也不要为了"快点看到功能"而跳过 - 骨架放好后,后续 Phase 的代码模型自然续写——因为"这个项目的代码长什么样"已经定了
**Continuous Development Mode**: Has code + has DEV-PLAN.md -> develop by Phase, **one Phase per /dev-builder invocation**. Each Phase: Plan Mode to plan implementation -> per-Task review + commit -> Phase four-step verification -> user confirmation -> **force stop** (suggests running **/code-review** for a holistic Phase-level review). User must call /dev-builder again for next Phase.
**Change-Scoped Mode**: Invoked from `/change-manager apply` with `change-name=<name>` -> read `changes/<name>/` (specs, design, tasks), execute **only** unchecked items in `changes/<name>/tasks.md`. Do not pull unrelated DEV-PLAN Phases. Still uses implementer + TDD + two-tier review per Task.
[Not For] - Fixing bugs in existing code -> use /bug-fixer instead - Reviewing code quality -> use /code-review instead - Planning development phases -> use /dev-planner instead - Gathering requirements -> use /product-spec-builder instead
[Dependency Check] Executed automatically as the first step when the Skill starts.
Required:
- Product-Spec.md -> if missing, prompt user to call /product-spec-builder first
- DEV-PLAN.md -> if missing, prompt user to call /dev-planner first
- All system tools and runtime environments listed in the DEV-PLAN tech stack table
Optional:
- Design-Brief.md -> if missing, mark as "no design specification mode"
- DESIGN.md -> if present, use frozen tokens for Tailwind/theme (priority over Brief for exact values)
- Design tool MCP -> if missing, mark as "no design draft mode"
- gh CLI -> if available, can automatically create GitHub repo and push
- playwright -> if available, can do UI automated testing
- **Dependency Graph** (`dep-graph`) -> if available, enables blast-radius analysis for impact assessment and risk-scored complexity gating
Installation Strategy:
- When required dependencies are missing or version requirements not met, the Agent autonomously determines the installation method and installs directly — no manual user operation needed
- If user permissions or interaction is needed, prompt the user to act
- When optional dependencies are missing, mark as degraded mode and continue working — do not block the workflow
[First Principles]
编码前必读 references/first-principles.md。TDD、implementer 隔离、验证即证据 — 非协商。
[Shared Discipline]
Karpathy 四原则 → ../_shared/karpathy-discipline.md(全文 core/docs/behavior-rules.md)
finding 的 action 三分(auto-fix/ask-user/no-op)驱动 Step 14.6 立即 escalate → ../_shared/finding-actions.md
[HARD-GATE]
主 Session MUST NOT 直接 Write/Edit 业务代码;每 Task dispatch implementer 并创建 .forge/implementer-session.json;首次改代码前 MUST git worktree。
**Exception 1 — UI / Integration Phase**: 主 session 可直接 Write/Edit,不强制 implementer 和 worktree。理由:UI 代码需要完整组件上下文,implementer 隔离得不偿失(Dogfood #2 验证)。
**Exception 2 — Small Phase (any Nature)**: 当 Phase 的 Key Files ≤3 且 Deliverables ≤5(即"Small Phase"),主 session 可直接 Write/Edit,不强制 implementer 和 worktree。理由:小型 Phase 的 implementer 冷启动开销 > 实际编码工作量(Dogfood #3 验证:Greenfield CLI 全部 5 Phase 均小到不值得隔离)。尺寸判定详见 `references/workflow.md` § Nature Gate Step 1.5。
What ships with it
13 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- commands/dev-builder.md 678 B
- references/anti-ai-slop-checklist.md 4.7 KB
- references/anti-rationalization.md 3.0 KB
- references/development-dimension-checklist.md 4.0 KB
- references/development-rules-checklist.md 10 KB
- references/development-strategies.md 5.4 KB
- references/first-principles.md 6.3 KB
- references/gc-audit-routing.md 2.4 KB
- references/phase-completion-assessment.md 9.9 KB
- references/sub-agent-isolation.md 4.9 KB
- references/workflow.md 32 KB
- references/zoom-out-pass.md 1.0 KB
- skill.json 926 B
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.
- 1mo ago First seen · 205 lines · 0 tokens per session scan A 33e93e786860
dev-builder is a skill published in the GitHub repository zxpmail/ReqForge (18 stars, last pushed 10d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 3,428 tokens. 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.
Other skills, from other repositories
chinese-git-workflow
A reference for configuring Git with Chinese code-hosting services such as Gitee, Coding.net, GitLab China, and CNB, including SSH, HTTPS, credentials, CI, and repository mirroring.
chinese-documentation
A Chinese technical-documentation style guide covering spacing, punctuation, numbers, terminology, and links when Chinese and English appear together.
requirements-writer
Use when converting PM intake into a slice SPEC.md with observable acceptance outcomes, explicit scope, and advisory sibling build dependencies.
orchestration-team
Use when coordinating assignments, selected review boundaries, or blocked work across a rig.
oversight-team
Use when you are a seat on the oversight pod (a standing monitor-mode rig that keeps OTHER rigs healthy), configuring or running the drift detectors, or choosing whether to intervene vs escalate. Covers the pull-not-poll posture, the v0 detectors (premature-park, process-drift, off-task, token-burn), the intervention…
oma-hwp
Convert Hangul HWP, HWPX, or HWPML documents to Markdown with kordoc. Use for text, table, and image extraction from these formats.