hyper-debug

A structured debugging guide for test failures, exceptions, regressions, and unexpected behavior. It uses a reproduce–hypothesize–experiment cycle before fixing the cause.

In plain words
What is it for?
Use it to reproduce a problem, state the expected and actual behavior, test one cause at a time, add a failing test, apply the fix, and verify it.
Why use it?
It reduces guess-and-check debugging and helps distinguish a confirmed cause from an assumption. It also calls for a regression test so the same bug is less likely to return.

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/zeikar/hyperclaude/hyper-debug
Any agent
npx skills add zeikar/hyperclaude --skill hyper-debug
Clone the repo
git clone --depth 1 https://github.com/zeikar/hyperclaude

Made for: Claude Code, Codex.

Per session 46 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 453 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 $0.00046 $0.00453
Opus 5 $0.00023 $0.00227
Sonnet 5 $0.00009 $0.00091
Haiku 4.5 $0.00005 $0.00045

Measured yesterday against content hash bfb685405363, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

hyper-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 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/hyper-debug/SKILL.md · 38 lines

What it actually says

hyper-debug

Debugging discipline tailored to this author's preferences.

Skip when

  • You already know the exact line and exact fix (typo, missing import, obvious null check). Just fix it.
  • The "bug" is a one-character correction that doesn't merit a hypothesis.

If you catch yourself rationalizing ("the cause is probably X, let me just try…") — that's hypothesis stacking. Run the cycle.

The cycle

  1. Reproduce. Smallest possible repro. If you can't reproduce, you can't debug.
  2. State the expected vs actual. Two sentences. "I expected X. I got Y."
  3. Hypothesize one cause. One. Not five. Pick the most likely.
  4. Predict what would prove the hypothesis right or wrong before running anything. "If hypothesis is correct, X should happen."
  5. Run the experiment. Compare to prediction.
  6. If wrong, hypothesize again. New hypothesis. Don't stack them.
  7. Once cause is confirmed, write a failing test that captures the bug before fixing — so it doesn't regress silently later.
  8. Fix. Run. Commit.

Anti-patterns

  • Hypothesis stacking — trying five "fixes" at once and not knowing which one worked.
  • Skipping the regression test — "the fix is obvious, no test needed." Write the test.
  • Logging-only debugging — endless console.log without forming hypotheses. Logs serve a hypothesis, not the other way around.
  • Reading code top-to-bottom — start at the failing assertion and walk backward, not forward.

When to escalate

If after 2-3 hypotheses you still don't know the cause, stop iterating. Re-read the error message slowly, check if your repro is actually exercising the buggy code, or take a 5-minute break — you have a wrong assumption somewhere upstream.

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 · 38 lines · 46 tokens per session scan A bfb685405363

Subscribe to this mod's changes

hyper-debug is a skill published in the GitHub repository zeikar/hyperclaude (3 stars, last pushed 14d ago), licensed MIT. It adds 46 tokens to every session and 453 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 2026-08-31.