debugging

Systematically debug code issues using proven methodologies. Use when encountering errors, unexpected behavior, or performance problems. Handles error analysis, root cause identification, debugging strategies, and fix verification.

Install with Tessl CLI

npx tessl i github:supercent-io/skills-template --skill debugging

What are skills?

1.09x

Quality

79%

Does it follow best practices?

Impact

78%

1.09x

Average score across 3 eval scenarios

Optimize this skill with Tessl

npx tessl skill review --optimize ./.agent-skills/debugging/SKILL.md

SKILL.md

Review

Evals

Evaluation results

71%

-2%

Log Analysis Pipeline: Out-of-Memory Failure

Memory leak diagnosis and fix

Criteria

Without context

With context

Memory profiler use

Root cause identified

100%

Generator-based fix

100%

Fix targets root cause

100%

Edge cases handled

100%

Regression test naming

100%

Regression test docstring

100%

Parametrized edge case tests

13%

Findings documented

83%

Without context: $0.9214 · 3m 22s · 34 turns · 41 in / 13,939 out tokens

With context: $0.7977 · 2m 36s · 34 turns · 153 in / 9,664 out tokens

90%

Concurrent Task Tracker: Intermittent Count Errors

Race condition identification and fix

Criteria

Without context

With context

Race condition identified

100%

threading.Lock added

100%

Lock used as context manager

100%

Reproduction test present

100%

Minimal test isolation

100%

DEBUG logging added

Fix does not break existing behavior

100%

Regression test docstring

100%

Findings documented

100%

Without context: $0.5937 · 2m 34s · 27 turns · 34 in / 9,154 out tokens

With context: $0.5239 · 1m 56s · 25 turns · 28 in / 6,644 out tokens

73%

22%

Order Processing Pipeline: Incorrect Total Calculation

Systematic bug isolation with divide-and-conquer

Criteria

Without context

With context

DEBUG print tracing

Divide-and-conquer isolation

26%

86%

Correct step identified

100%

Root cause described

100%

Fix is correct

100%

Regression test naming

70%

100%

Regression test docstring

100%

Parametrized edge cases

33%

One-change-at-a-time evidence

28%

100%

Without context: $0.4982 · 1m 48s · 22 turns · 29 in / 7,352 out tokens

With context: $0.6926 · 2m 26s · 29 turns · 459 in / 9,302 out tokens

Evaluated: 3 days ago
Agent: Claude Code
Model: Claude Sonnet 4.6

Table of Contents

Log Analysis Pipeline: Out-of-Memory Failure Concurrent Task Tracker: Intermittent Count Errors Order Processing Pipeline: Incorrect Total Calculation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.