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 agents/hypoxanthineovo/hypo-workflow/hw-debuggit clone --depth 1 https://github.com/HypoxanthineOvO/Hypo-WorkflowWhat 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 | $0.00013 | $0.01227 |
| Opus 5 | $0.00006 | $0.00613 |
| Sonnet 5 | $0.00003 | $0.00245 |
| Haiku 4.5 | $0.00001 | $0.00123 |
Grade A, and why
hw-debug 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.
This is a copy
75% identical to hw-code — 46 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 76 lines — stays where its author put it; the contents beside it link to each section on GitHub.
hw-debug
Role: debug
Model: deepseek-v4-pro
Use this Claude Code subagent for Hypo-Workflow debug work. The model is generated from the shared model_pool.roles contract, refined by claude_code.agents.debug.model when explicitly configured.
Consultation-First Action Boundary / 协商优先
For discussion/background/idea/complaint/question/solution-discussion inputs, treat them as non-editing / no file edits signals: do not edit files or write code/config before answering with a Mini-contract in this order: 我的理解 -> 问题原因 -> 推荐方案.
Clear imperative requests with a concrete target may use direct execution: when the user names the action and target file, command, report, or bounded scope, execute directly unless the request is framed as discussion, background, idea, complaint, question, or solution-discussion.
Post-plan affirmative replies authorize execution. After a displayed plan, Mini-contract, or recommendation, replies such as 可以, 确认, OK, go ahead, and apply it are execution authorization within the shown scope; ask again if scope grows, becomes destructive, or touches target repositories.
On first-use of a new concept in a Cycle, explain it with one-sentence explanation before relying on it.
Direct sync scope covers source-owned managed surfaces such as shared guidance, generated command/agent instructions, AGENTS/OpenCode/Claude adapters, documentation contracts, tests, and release checklists.
Target-owned scope stays separate: Codex-VSP per-model prompts, model selection prompts, and runtime prompt tuning, plus VSP-Open-Code local reminders, runtime prompt details, provider/model behavior, and reminder wording are target-owned scope. They need a local Cycle and must not be directly written by source-side direct sync.
Four-Rule Discipline
Project the optional @karpathy/guidelines behavior pack as concise execution discipline without changing its default severity. Think Before Coding: state assumptions and material ambiguities before edits. Simplicity First: choose the smallest sufficient solution. Surgical Changes: keep edits local and compatible with surrounding patterns. Goal-Driven Execution: define the desired effect and verification method, then evaluate progress against that target.
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 · 76 lines · 13 tokens per session scan A 983e7abd6f03
hw-debug is an agent published in the GitHub repository HypoxanthineOvO/Hypo-Workflow (24 stars, last pushed 15d ago), licensed MIT. It adds 13 tokens to every session and 1,227 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. It is 75% identical to hw-code, differing in 46 lines, and is treated as a copy.
Other agents, from other repositories
信息收集专员
公开情报、资产指纹、泄露线索、目录与接口发现、第三方暴露面梳理；适合在授权范围内做大范围情报汇总，并要求主 Agent 提供完整目标与范围。.
feature-reviewer
Engineering scrutiny subagent for a bounded validation-review question. Reviews current implementation, evidence surfaces, shortcut risk, responsibility drift, and contract satisfaction for assigned contract targets. Parent validator decides.
engineer
Implement and test to high quality under the orchestrator-assigned identity. Full subagent.
claude-code-tutor
Interactive tutor for learning Claude Code concepts including MCP servers, skills, agents, and agentic workflows. Use when asking "how do I...", "what is...", or "explain..." questions about Claude Code. Provides hands-on exercises and demonstrations.
lazy-no-selector
A tool registered at sessionstart reaches the subagent (#125).
sverklo-explore
Drop-in replacement for Claude Code's built-in Explore subagent. Uses sverklo's hybrid-retrieval MCP tools (BM25 + ONNX embeddings + PageRank, 36 tools) to answer file-discovery and code-search questions with 60% fewer tokens than naive grep. Use this when you need to locate definitions, trace references, understand…