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 skills/pieceoffall/jdbg/mcpnpx skills add PieceOfFall/jdbg --skill mcpgit clone --depth 1 https://github.com/PieceOfFall/jdbgWrote 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/pieceoffall/jdbg/mcp)<a href="https://agentmods.dev/skills/pieceoffall/jdbg/mcp"><img src="https://agentmods.dev/badge/skills/pieceoffall/jdbg/mcp.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 5.1 | $0.00105 | $0.06993 |
| Opus 5 | $0.00053 | $0.03497 |
| Sonnet 5 | $0.00021 | $0.01399 |
| Haiku 4.5 | $0.00011 | $0.00699 |
Grade A, and why
jdbg 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 — 367 lines — stays where its author put it; the contents beside it link to each section on GitHub.
jdbg — interactive Java debugging for agents
jdbg is an MCP server that drives the JDK's jdb debugger as a stateful background session, so you
can debug a Java program across many separate tool calls. Each tool call (launch, break_at, run,
locals, …) sends ONE command to the session and returns a structured result. A background daemon keeps
the JVM and debugger state alive between calls — set a breakpoint in one call, run in the next, inspect in the
next.
The tools appear as
mcp__jdbg__<tool>(e.g.mcp__jdbg__break_at). Below they are named bare (break_at) for brevity. The first tool call auto-starts the background daemon — no setup step is needed.
Core loop: react to each result. Call one tool, read its structured output, then decide the next call.
Do NOT batch commands or tune timing — jdbg waits for jdb to be ready (no sleeps), and a hung program
returns a non-destructive Timeout, not a freeze.
When to use
- Find the real runtime value of a variable / field / expression at a line (not by reading code).
- Diagnose why an exception or NPE is thrown — the stack and locals at the throw site.
- Inspect threads; find what a blocked or deadlocked thread is waiting on.
- Step through execution line by line to see actual control flow.
- Attach to an already-running JVM that has JDWP enabled.
When NOT: logic simple enough to read; a compile/build error (that is javac, not a debugger); no
running JVM and nothing launchable.
Start a session — two ways
Launch a program under the debugger (you have the main class + classpath) — call launch with
main_class (and usually classpath; pass program args via app_args):
launch { "main_class": "com.example.Main", "classpath": "out", "sourcepath": "src", "app_args": ["arg1", "arg2"] }
Returns a session id, state loaded (JVM not started yet). Set breakpoints, then call run.
Attach to a running JVM started with JDWP — call attach:
attach { "host": "localhost", "port": 5005, "sourcepath": "src" }
The default JDI backend returns state running for attach; set breakpoints, then call cont (attach has no
run). If local JDI prerequisites are missing, an omitted backend automatically falls back to jdb and
returns a note; explicit backend: "jdi" means JDI is required and does not fall back.
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 · 367 lines · 105 tokens per session scan A 547b8431449a
jdbg is a skill published in the GitHub repository PieceOfFall/jdbg (11 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 105 tokens to every session and 6,993 once invoked, about $0.0005 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.
Other skills, from other repositories
java-debug
Debug a live Java application via the jdwp-inspector MCP server — breakpoints, runtime state, expression eval, variable mutation, exception-throw-site catches, non-intrusive line/exception logpoints, and field watchpoints (suspend or log on every read/write of a specific field).
release
Use when cutting a new release of the jdwp-debugging plugin — bumping the user-facing version, writing a CHANGELOG entry, committing, tagging, and pushing. Triggers — "do a release", "ship X.Y.Z", "release the last commits", "bump the version", "tag a release".
debug-live
Guides root-cause investigations with debugging capabilities by setting breakpoints, starting a debug session, stepping through execution, inspecting variables, and tracing symptoms back to their origin. Prefer it for runtime bugs, failing tests, exceptions, crashes, hangs, wrong/null values, and unexpected output…
flask-werkzeug-attack
Exploit Flask/Werkzeug debugger exposure for traceback and SECRET leaks.
embedded-debugging
Debug embedded firmware on real hardware through the jlink-mcp server — crashes, hangs, peripheral misconfiguration, silent devices. Covers loading ELF symbols for source-level backtraces, supplying an SVD for named peripheral fields, and the halt/read/resume discipline. Use whenever debugging a microcontroller over a…
dbg
Debug applications using the dbg CLI debugger. Supports Node.js (V8/CDP), Bun (WebKit/JSC), Python (debugpy/DAP), Java (JDWP/DAP), and native code via LLDB (DAP). Use when: (1) investigating runtime bugs by stepping through code, (2) inspecting variable values at specific execution points, (3) setting breakpoints and…