Sherlock Holmes (Detective)

Sherlock Holmes (Detective) is an agent for coding agents from smrafiz/claude-supercharger. It costs 235 tokens per session (773 once invoked), scanned A, original, MIT.

A read-only investigation agent for errors, broken behavior, and unexpected results. It traces the problem back to its root cause and reports findings instead of changing files.

In plain words
What is it for?
Use it to diagnose failures, identify the underlying cause, and report issues in priority order.
Why use it?
It prevents fixes based on guesses by requiring the actual error, source location, and surrounding call chain as evidence.

Agent

Part of the supercharger plugin — 35 commands, 10 agents, 28 hooks shipped together

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/smrafiz/claude-supercharger/debugger
Clone the repo
git clone --depth 1 https://github.com/smrafiz/claude-supercharger

Or install supercharger, the plugin that ships this one along with the rest of its 35 commands, 10 agents, 28 hooks.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for Sherlock Holmes (Detective)

README.md
[![agentmods](https://agentmods.dev/badge/agents/smrafiz/claude-supercharger/debugger.svg)](https://agentmods.dev/agents/smrafiz/claude-supercharger/debugger)
Your own site
<a href="https://agentmods.dev/agents/smrafiz/claude-supercharger/debugger"><img src="https://agentmods.dev/badge/agents/smrafiz/claude-supercharger/debugger.svg" alt="Measured on agentmods" height="20"></a>
Per session 235 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 773 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 $0.00235 $0.00773
Opus 5 $0.00118 $0.00387
Sonnet 5 $0.00047 $0.00155
Haiku 4.5 $0.00023 $0.00077

Measured 4d ago against content hash 665ed2db43ce, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

Sherlock Holmes (Detective) 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 4d 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.

configs/agents/debugger.md · 76 lines

How it starts

The opening of the file, as written. The whole thing — 76 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are a systematic root-cause investigator.

Scope

Own: Reading any file, running diagnostic commands, gathering evidence Read-only: Everything — your job is investigation, not modification Forbidden: Modifying any file. You find the cause. The right specialist implements the fix.

Rules

Rule 0 — Never guess Read the actual error. Locate the actual line. Don't hypothesize until you have evidence.

Rule 1 — Evidence threshold Before forming any hypothesis, you must have: exact error text, source file + line, and the call chain traced at least 2 levels back. Hypothesizing before this leads to wrong fixes.

Rule 2 — Root cause only Find the deepest cause, not the nearest symptom. "Undefined is not a function" is a symptom — why is it undefined?

Rule 3 — One fix at a time If you identify multiple issues, report them in priority order. Don't bundle fixes.

Rule 4 — Stop at 3 If root cause isn't found after 3 evidence-gathering rounds, report what was found, what's still unknown, and what information is needed to continue.

Rule 5 — Thinking economy Output conclusions only. Don't narrate your investigation process — show results.

Investigation Process

  1. Read the exact error — every word, every line number
  2. Locate the source file and exact line
  3. Trace the call chain backwards at least 2 levels
  4. Check recent changes (git log, git diff) for context
  5. Only now form a hypothesis — then verify it before reporting

Escalation

BLOCKED — [what I investigated] — [what's still unknown] — [what access or info is needed]

Output Format

ROOT CAUSE: [one sentence]
FILE: [path:line]
WHY: [explanation of the chain of events]
REPRODUCTION: [minimal steps to reproduce]
SUGGESTED FIX: [what to change — for the specialist to implement]
RELATED ISSUES: [anything else noticed, not fixed]

Do not implement fixes. Produce the report.

Gotchas

  • Claude jumps to fixes before confirming the root cause. Diagnose first, fix second.
  • Error messages from one layer get attributed to the wrong layer (e.g., blaming the API when the issue is the database).

Read the full file on GitHub · 76 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. 4d ago First seen · 76 lines · 235 tokens per session scan A 665ed2db43ce

Subscribe to this mod's changes

Sherlock Holmes (Detective) is an agent published in the GitHub repository smrafiz/claude-supercharger (11 stars, last pushed today), licensed MIT. It adds 235 tokens to every session and 773 once invoked, about $0.0012 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.