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 skills add decodingai-magazine/building-a-coding-agent-from-scratch-course --skill demo-2-bug-huntgit clone --depth 1 https://github.com/decodingai-magazine/building-a-coding-agent-from-scratch-courseWrote 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/skills/decodingai-magazine/building-a-coding-agent-from-scratch-course/demo-2-bug-hunt)<a href="https://agentmods.dev/skills/decodingai-magazine/building-a-coding-agent-from-scratch-course/demo-2-bug-hunt"><img src="https://agentmods.dev/badge/skills/decodingai-magazine/building-a-coding-agent-from-scratch-course/demo-2-bug-hunt/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/skills/decodingai-magazine/building-a-coding-agent-from-scratch-course/demo-2-bug-hunt"><img src="https://agentmods.dev/badge/skills/decodingai-magazine/building-a-coding-agent-from-scratch-course/demo-2-bug-hunt.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 | $0.00038 | $0.00621 |
| Opus 5 | $0.00019 | $0.00311 |
| Sonnet 5 | $0.00008 | $0.00124 |
| Haiku 4.5 | $0.00004 | $0.00062 |
Grade C, and why
demo-2-bug-hunt scanned grade C with 1 finding 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 10d 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.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
`rm -rf .decode/outputs/bug-hunt && mkdir -p .decode/outputs && cp -r .decode/skills/demo-2-bug-hunt/references/buggy_repo .decode/outputs/bug-hunt` How it starts
The opening of the file, as written. The whole thing — 51 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Play detective on a small statistics package: reproduce the failures, hunt down the bugs, fix them until every test passes, and file a case report.
Setup
- Copy the seeded project into a working directory you own (fresh — delete a stale one first):
rm -rf .decode/outputs/bug-hunt && mkdir -p .decode/outputs && cp -r .decode/skills/demo-2-bug-hunt/references/buggy_repo .decode/outputs/bug-hunt - Change into it and run the suite to see the failures:
cd .decode/outputs/bug-hunt/ && uv run pytest -q - Save the crime scene: keep the exact failing-test output — the failure names and the expected-vs-actual values — you will quote it in the case file at the end.
Exactly two tests fail as committed — one in median, one in variance. Do NOT edit the tests;
they encode the correct behaviour. Fix stats.py instead.
Hunt
Track the investigation with todo_write (reproduce → suspect A → suspect B → verify → case
file) and tick items off as you close them.
- Read
test_stats.pyto learn the contract each failing test asserts (the expected values are the spec). - Read
stats.pyand locate the defect behind each failure withgrep/read. The LSP diagnostics that surface on your edits will help you catch typos and type slips as you go. - There are two independent bugs:
medianreturns the wrong element for odd-length inputs (an indexing off-by-one).variancecomes back with the wrong sign (it should never be negative).
Fix and verify
- Fix the root cause of each bug in
stats.py— the smallest correct change, not a special case that only satisfies the one test input. - Re-run
uv run pytest -qand confirm the whole suite is green. Keep that green one-liner too.
File the case report
Write .decode/outputs/bug-hunt/CASE_FILE.md — short, punchy, detective-flavoured:
- The symptoms — the failing test names and the quoted expected-vs-actual output from setup step 3.
- The culprits — one section per bug: the guilty line (
stats.py:<line>), the root cause in one sentence, and the before/after of the fixed line. - Case closed — the quoted green
pytest -qsummary line proving the suite passes.
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 10d ago First seen · 51 lines · 38 tokens per session scan C 5c967167f0e2
demo-2-bug-hunt is a skill published in the GitHub repository decodingai-magazine/building-a-coding-agent-from-scratch-course (374 stars, last pushed today), licensed Apache-2.0. It adds 38 tokens to every session and 621 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
openocd
A toolkit for programming and debugging microcontrollers through hardware probes such as ST-Link or CMSIS-DAP. OpenOCD can act as a GDB Server, allowing a debugger to inspect source code, registers, memory, breakpoints, and program execution.
serial
A tool for communicating with devices over serial ports, such as USB-to-UART connections. It can find ports, watch incoming data, send messages, save logs, and display raw bytes in hexadecimal form.
gcc
A build tool for embedded CMake projects using arm-none-eabi-gcc, a compiler that turns code into firmware for ARM microcontrollers. It supports CMake presets, which are saved build configurations.
net
A toolkit for diagnosing network connections and traffic in embedded systems. A packet capture is a recorded sample of network traffic, and this tool can create or analyze captures in formats such as pcap and pcapng.
probe-rs
A command-line tool wrapper for probe-rs, which connects a computer to a microcontroller through a hardware debug probe. It supports common firmware programming and debugging operations over SWD or JTAG.
terminal
An interactive terminal session that keeps a live connection to a local shell, an SSH remote computer, or a serial device. It lets you send commands and read the next output while preserving the session’s login, menu, or program state.