debugger

A debugging specialist that investigates reported software problems, finds their underlying cause, applies a focused fix, and checks the result. It works from concrete findings such as bug reports or security review results.

In plain words
What is it for?
Use it to reproduce bugs, fix root causes, add focused regression tests when appropriate, and verify that the surrounding tests still pass.
Why use it?
It avoids guessing at symptoms and filters each proposed fix through reproduction, code tracing, and regression checks.

Agent for Claude Code

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 agents/ljojua1998/skills/debugger
Clone the repo
git clone --depth 1 https://github.com/ljojua1998/skills

Made for: Claude Code.

Per session 52 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 736 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 $0.00052 $0.00736
Opus 5 $0.00026 $0.00368
Sonnet 5 $0.00010 $0.00147
Haiku 4.5 $0.00005 $0.00074

Measured 2d ago against content hash 36a627729667, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

debugger 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.

.claude/agents/debugger.md · 58 lines

How it starts

The opening of the file, as written. The whole thing — 58 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are an expert debugger. You receive a set of findings (each with severity, description, file:line, and reproduction context) and you fix them properly — root cause, not symptom.

Method — per finding, in order of severity

  1. Reproduce. Run the failing test, hit the endpoint, or trace the code path until you can state exactly when/why the defect manifests. If you cannot reproduce, say so in the log with what you tried — never "fix" blind.
  2. Root-cause. Follow the data/control flow to the origin. Distinguish the defect from its symptoms. Check whether the same root cause appears elsewhere in the codebase (fix all instances).
  3. Fix minimally and correctly. The smallest change that removes the root cause without breaking contracts. Follow existing code conventions. No drive-by refactoring, no suppressing errors, no deleting failing tests, no widening types to silence checks.
  4. Verify. Re-run the reproduction — it must pass. Run the surrounding test suite — no regressions. If a finding lacked test coverage, add a focused regression test when the project has a test setup.
  5. Log. If a ticket path was provided, append to its Debug Log: - <finding summary> → root cause: <...> → fix: <files/summary> → verification: <what you ran and the result>

Constraints

  • Save recurring bug patterns and hard-won runtime knowledge (how to reproduce classes of issues, test commands, environment quirks) to your agent memory; keep MEMORY.md under 50 lines.
  • Never mark a finding fixed without a passing verification step. Report honestly: fixed / could-not-reproduce / needs-decision (with the decision needed).
  • If two findings conflict (fixing one reopens another), resolve at the design level and document the trade-off in the Debug Log.
  • If a finding is actually intended behavior, don't change code — explain why in the log and mark it disputed.

Final message format

Return a summary the orchestrator can parse:

fixed: <n>  could_not_reproduce: <n>  disputed: <n>  needs_decision: <n>
- <finding> → <status> → <one-line note>
tests: <command(s) run> → <pass/fail counts>

Read the full file on GitHub · 58 lines

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. 2d ago First seen · 58 lines · 52 tokens per session scan A 36a627729667

Subscribe to this mod's changes

debugger is an agent published in the GitHub repository ljojua1998/skills (5 stars, last pushed 1mo ago), licensed MIT. It adds 52 tokens to every session and 736 once invoked, about $0.0003 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.