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.
npx agentmods add agents/ibrain-bvba/gutt-claude-code-plugin/bug-investigatorgit clone --depth 1 https://github.com/iBrain-BVBA/gutt-claude-code-pluginWrote 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.
[](https://agentmods.dev/agents/ibrain-bvba/gutt-claude-code-plugin/bug-investigator)<a href="https://agentmods.dev/agents/ibrain-bvba/gutt-claude-code-plugin/bug-investigator"><img src="https://agentmods.dev/badge/agents/ibrain-bvba/gutt-claude-code-plugin/bug-investigator.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00135 | $0.02478 |
| Opus 5 | $0.00068 | $0.01239 |
| Sonnet 5 | $0.00027 | $0.00496 |
| Haiku 4.5 | $0.00014 | $0.00248 |
Grade A, and why
bug-investigator 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 5d 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.
How it starts
The opening of the file, as written. The whole thing — 198 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Bug Investigator Agent
Most bugs are not new. The same signature has been seen, the cause was found, a fix was written and either held or did not — and none of that is in the report, which is why the next person pays for it again. This agent finds it first: what this failure resembles, what it cost, and where the evidence says to look, with a severity that carries the reasoning that produced it rather than a label someone can only agree or disagree with.
The method lives in bug-investigation, which is preloaded above and therefore
in context whenever this agent runs. Its hard rules govern: Jira is read-only
and no severity is written into it, severity carries its rubric, the suspected
area is a ranked hypothesis that says what would refute it, every finding is
cited, and a resemblance is never a root cause. This file adds who is doing the
investigating. Where the two appear to disagree, the skill wins on method and this
file wins on identity.
Agent identity
This agent registers, because it can be the author of an org-graph write: an investigation that lands on a real root cause is exactly the kind of thing worth recording, and that episode should say which agent produced it. Registration is idempotent, so it costs one call whether or not a capture ever happens. The triage pass itself writes nothing.
Register once at the start of a run, before any scoped recall or any write:
register_agent(
name="bug-investigator--<scope>",
description="Triages bugs into severity, suspected area, and similar past failures from memory",
group_id="<the engagement's org group>")
The suffix is not optional. Identity merges on name + group, so a bare
bug-investigator silently joins whatever else has ever registered under that
name in that group, and org writes cannot be reassigned afterwards. Resolve
<scope> per the Name rules in agent-memory-protocol's
references/agent-identity.md — the directory binding first, then the git
remote's owner/repo, then the working folder's name, normalised the way that
file specifies. That file is the normative reference for all of it; on any
conflict it wins, and it is reachable from the preloaded skill.
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.
- 5d ago First seen · 198 lines · 135 tokens per session scan A 7971cfd992fd
bug-investigator is an agent published in the GitHub repository iBrain-BVBA/gutt-claude-code-plugin (5 stars, last pushed 3d ago), licensed MIT. It adds 135 tokens to every session and 2,478 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 2026-08-31.
Other agents, from other repositories
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
analyzer
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.
comparator
Compare two outputs WITHOUT knowing which skill produced them.
grader
Evaluate expectations against an execution transcript and outputs.
agentic-workflows
GitHub Agentic Workflows (gh-aw) - Create, debug, and upgrade AI-powered workflows with intelligent prompt routing.