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.
npx agentmods add skills/magebyte-zero/spec-superflow/need-explorernpx skills add MageByte-Zero/spec-superflow --skill need-explorergit clone --depth 1 https://github.com/MageByte-Zero/spec-superflowWhat 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 | $0.00044 | $0.00928 |
| Opus 5 | $0.00022 | $0.00464 |
| Sonnet 5 | $0.00009 | $0.00186 |
| Haiku 4.5 | $0.00004 | $0.00093 |
Grade A, and why
need-explorer 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.
How it starts
The opening of the file, as written. The whole thing — 107 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Need Explorer
Turn a rough idea into a stable change definition before writing artifacts.
Primary Goal
Agree on: problem, scope, non-goals, success criteria, whether to split before specification.
Process
1. Inspect Context First
Before asking questions, understand what exists and what constraints are in place.
2. One Question at a Time
Ask a single clear question, wait for the answer, digest, then ask the next. Never ask 3+ questions at once. Each answer informs the next question.
3. Prefer Multiple-Choice Questions
Present 2-3 options when reasonable answers are finite. This reduces cognitive load and surfaces unconsidered choices.
4. Propose 2-3 Approaches with Trade-Offs
For each approach: what it is, upside, downside, best-for. Then recommend one and explain why. Never present a single path — always name at least one alternative.
5. Validate Before Concluding
Restate what you heard: "Here's what I'm hearing: [problem, scope, non-goals, success criteria]. Does this match?" Incorporate corrections and re-validate.
6. DP-1: Requirement Confirmation Gate
After user confirms the summary:
ssf state set <change-dir> dp_1_result "confirmed: <one-line summary>"
ssf state set <change-dir> dp_1_timestamp $(date -u +%Y-%m-%dT%H:%M:%SZ)
DP-1 confirms scope, non-goals, and success criteria before artifact creation.
7. Hand Off
Once DP-1 is recorded, hand off to spec-writer.
Anti-Patterns
- Skipping exploration: "Simple" changes have scope too. Five minutes of exploration prevents two hours of rework.
- Proposing solutions before clarifying: If the user says "add caching," first ask what problem caching solves.
- Exploring indefinitely: Stop when change name, problem statement, scope, non-goals, success criteria, and decomposition decision are all clear.
Exploration Standard
You must leave exploration with: a usable change name, a crisp problem statement, scope boundaries, non-goals, success criteria, and a decomposition decision (one change or split).
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.
- 2d ago First seen · 107 lines · 44 tokens per session scan A 8d4f8d5c0179
need-explorer is a skill published in the GitHub repository MageByte-Zero/spec-superflow (776 stars, last pushed 4d ago), licensed MIT. It adds 44 tokens to every session and 928 once invoked, about $0.0002 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.
Other skills, from other repositories
chinese-documentation
中文文档排版参考——中英文空格、全半角标点、术语保留、链接格式、中文文案排版指北约定。仅在用户显式 /chinese-documentation 时调用，不要根据上下文自动触发。.
chinese-git-workflow
国内 Git 平台配置参考——Gitee、Coding.net、极狐 GitLab、CNB 的 SSH/HTTPS/凭据/CI 接入差异与镜像同步配置。仅在用户显式 /chinese-git-workflow 时调用，不要根据上下文自动触发。.
chinese-commit-conventions
中文 commit 与 changelog 配置参考——Conventional Commits 中文适配、commitlint/husky/commitizen 中文模板、conventional-changelog 中文配置。仅在用户显式 /chinese-commit-conventions 时调用，不要根据上下文自动触发。.
systematic-debugging
Skill "systematic-debugging" from jnMetaCode/superpowers-zh, covering 系统化调试, 概述, 铁律, 何时使用 and 四个阶段.
workflow-runner
在 Claude Code / OpenClaw / Cursor 中直接运行 agency-orchestrator YAML 工作流——无需 API key，使用当前会话的 LLM 作为执行引擎。当用户提供 .yaml 工作流文件或要求多角色协作完成任务时触发。.
test
Proves that a code change meets its acceptance criteria. Uses focused automated checks and a real browser for browser-facing work. Use to test or verify a diff, branch, PR, URL, or user flow.