bug-hunter
SkillDev toolsLets your agent systematically find and fix bugs by tracing symptoms to root causes.
Available today. Use it from your connected AI after setup.
No other account needed.
Connect ahel once, and every AI you use reads what you have installed.
Then ask your AI: use the bug-hunter skill
About this capability
Searches code specifically for potential bugs, edge cases, race conditions, and incorrect assumptions rather than general style/quality. Use when the goal is finding what's broken, not improving what already works.
What this skill tells your AI
The instructions your AI receives, as published by codebygarv/ai-skills in skills/development/bug-hunter/SKILL.md and read by ahel’s review.
Purpose
Hunt specifically for defects — not style, not maintainability, not architecture. The single question driving this skill is: "under what real input or timing, does this code produce a wrong result or crash?"
When to Use
- The user wants bugs found, specifically — not a general quality review.
- Before shipping code that handles money, user data, or concurrency, where a missed bug is expensive.
- Debugging: the user suspects a bug exists somewhere in a file/module but hasn't found it yet.
What to Analyze
- Boundary conditions — empty collections, zero, negative numbers, max values, single-element cases.
- Null/undefined/missing data — every place data is accessed, ask what happens if it's absent.
- Type coercion and comparison bugs — loose equality, implicit conversions, unit mismatches (ms vs. seconds, cents vs. dollars).
- Concurrency/race conditions — shared mutable state, async operations that assume ordering that isn't guaranteed, double-submit/double-click scenarios.
- Off-by-one and loop errors — inclusive/exclusive bounds, iterator invalidation.
- Incorrect assumptions — code that assumes an invariant (sorted input, unique IDs, non-empty list) that isn't actually enforced anywhere.
Output Format
- Each bug as its own entry: file/line, trigger condition (the specific input/timing that causes it), actual vs. expected behavior, suggested fix.
- Ordered by severity — data corruption/crash first, cosmetic/rare-edge-case issues last.
- No entry without a concrete triggering scenario — "this looks risky" without a scenario isn't a finding, it's a hunch; say so separately if worth a mention.
Avoid
- Reporting style or readability issues — that's Code Reviewer's job, not this skill's.
- Vague findings without a reproducing scenario.
- Flagging theoretical issues that the surrounding code already guards against (check the full context first).
Signals
- GitHub stars
- 25
- Forks
- 1
- Last commit
- Aug 2026
Advanced
- Catalog kind
- skill
- Gateway key
bug-hunter- Source
- github.com/codebygarv/ai-skills