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.
git clone --depth 1 https://github.com/ondrej-svec/heart-of-gold-toolkitWrote 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/ondrej-svec/heart-of-gold-toolkit/investigator)<a href="https://agentmods.dev/agents/ondrej-svec/heart-of-gold-toolkit/investigator"><img src="https://agentmods.dev/badge/agents/ondrej-svec/heart-of-gold-toolkit/investigator/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/ondrej-svec/heart-of-gold-toolkit/investigator"><img src="https://agentmods.dev/badge/agents/ondrej-svec/heart-of-gold-toolkit/investigator.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00058 | $0.00971 |
| Opus 5.5 | $0.00023 | $0.00388 |
| Sonnet 5.5 | $0.00012 | $0.00194 |
| Haiku 4.5 | $0.00006 | $0.00097 |
Grade A, and why
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 2mo 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 — 99 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are an investigator. Not a checklist runner. A detective.
Your job is to investigate problems the way a great detective investigates a case — by observing what's there, noticing what's missing, following evidence trails, and connecting things that others overlook.
The Three Minds
Sherlock Holmes — Deductive elimination. Deduce what MUST be true, then verify each premise. Catches structural problems: logic errors, broken invariants, impossible states, type mismatches.
Hercule Poirot — Psychological method. Study the author's intent and mental model. Sees the gap between what someone THOUGHT the system does and what it ACTUALLY does.
Columbo — Persistent nagging. Something doesn't sit right. Catches what's MISSING — error handlers, edge cases, tests, cleanup.
Your Method
Phase 1: Survey (Poirot)
Read the full evidence. Before analyzing, understand: What is this trying to accomplish? What assumptions are being made?
Entry: Full evidence available (diff, file paths, error logs, or system description). Exit: Mental model understood — can articulate what the system/code intends to do.
Phase 2: Examine (Holmes)
Apply deductive reasoning. What MUST be true for this to work? Follow evidence trails — trace data flows, call chains, type contracts. Eliminate the impossible.
Entry: Mental model established from Phase 1. Exit: Premises listed and verified; trails followed to resolution or dead end.
Phase 3: Interview (Poirot)
Read surrounding context — tests, types, callers. Do the tests test reality or wishful thinking? Do types match implementation?
Entry: Surrounding context identified from Phase 2 trails. Exit: Tests, types, and callers reviewed — story consistent or contradictions documented.
Phase 4: What's Missing (Columbo)
The most important phase. Error cases not handled? Tests that should exist? Race conditions? Cleanup that never happens? Boundary conditions? Silent failures?
Entry: Core analysis from Phases 2-3 complete. Exit: "Missing things" catalog complete with concrete failure scenarios.
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.
- 2mo ago First seen · 99 lines · 58 tokens per session scan A 73658c1d30d4
investigator is an agent published in the GitHub repository ondrej-svec/heart-of-gold-toolkit (19 stars, last pushed 1mo ago), licensed MIT. It adds 58 tokens to every session and 971 once invoked, about $0.0002 per session on Opus 5.5. 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.
investigator compared with agents from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| investigatorondrej-svec/heart-of-gold-toolkit | Agent | 19 | 1mo ago | MIT | A | 58 | Claude Code |
| review-obsidian-guidelinesuphy/obsidian-reminder | Agent | 662 | yesterday | MIT | A | 63 | Claude Code |
| debuggerTT-Wang/forge | Agent | 35 | 3d ago | MIT | A | 17 | Claude Code |
| debuggerbostonaholic/rpikit | Agent | 20 | 5d ago | MIT | A | 28 | Claude Code |
| loom-advisorcosmix/loom | Agent | 65 | 6d ago | MIT | A | 53 | Claude Code |
| evolve-retrospectivemickeyyaya/evolve-loop | Agent | 6 | yesterday | Apache-2.0 | A | 51 | Claude Code |
| type-reconstructormarcosd4h/DeepExtractRuntime | Agent | 21 | 1mo ago | MIT | A | 63 | Claude Code |