ai-debug

A debugging and merge-conflict resolution guide for software projects. It requires identifying the specific cause of a failure and writing a check that demonstrates it before changing the code.

In plain words
What is it for?
It is for broken behavior, regressions, errors, failing continuous-integration checks, and merge or rebase conflicts.
Why use it?
It prevents fixes based only on symptoms or guesses. It gives developers a reproducible failure, a file-and-line explanation, and verification that the change solved the underlying problem.

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/arcasilesgroup/ai-engineering/ai-debug
Any agent
npx skills add arcasilesgroup/ai-engineering --skill ai-debug
Clone the repo
git clone --depth 1 https://github.com/arcasilesgroup/ai-engineering

Made for: Claude Code, Codex.

Per session 133 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 646 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 $0.00133 $0.00646
Opus 5 $0.00067 $0.00323
Sonnet 5 $0.00027 $0.00129
Haiku 4.5 $0.00013 $0.00065

Measured 3d ago against content hash f62c9b81b83a, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

ai-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 3d 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.

.agents/skills/ai-debug/SKILL.md · 55 lines

What it actually says

Find the cause, not the symptom

What it produces

A named cause at file:line, a check that fails because of it, and only then a fix.

Steps

  1. Reproduce it. If you cannot reproduce it, say so plainly and stop guessing: the next useful thing is a way to reproduce it, not a change.
  2. Read the failing output in full. The first error is usually the real one and the rest are its consequences; the last error is the one people paste.
  3. Name a cause you can point at. file:line, and one sentence on why that line produces this symptom. "Probably a race" is not a cause. If two causes are plausible, say which observation would tell them apart, then go and make that observation.
  4. Before the fix, write the check that fails for this reason. A fix with no failing check before it is a change with an opinion attached.
  5. Fix the cause, at the place all the callers go through. Patching the one path the report named leaves every sibling caller broken, and the shared fix is usually the smaller diff.
  6. Run the check. Then run the suite. Then say what you changed and why it fixes the cause you named, not the symptom that was reported.
  7. If you are two attempts in and it is still not fixed, stop and say so. That is a rule, not a suggestion: the third attempt is where the guessing starts.

Conflicts

Read both sides for intent before touching either. Lock files and generated files are regenerated, never merged by hand. Migrations are ordered, not combined. If two people meant different things, that is a conversation, not a resolution.

What this is not

  • "I know what the bug is even though I cannot reproduce it" — a cause you cannot reproduce is a guess: the next useful thing is a way to reproduce it, not a change.

Done when

  • The cause is named at file:line and a person could disagree with it.
  • A check exists that fails without the fix and passes with it.
  • You said what you changed, in a sentence somebody could act on.
Files

What ships with it

1 file 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.

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. 3d ago First seen · 55 lines · 133 tokens per session scan A f62c9b81b83a

Subscribe to this mod's changes

ai-debug is a skill published in the GitHub repository arcasilesgroup/ai-engineering (54 stars, last pushed 4d ago), licensed Apache-2.0. It adds 133 tokens to every session and 646 once invoked, about $0.0007 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.