Borrowing it
Nothing to install: this file belongs to glslang/windbg-mcp. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/glslang/windbg-mcp/main/.claude/skills/eval-bench/SKILL.mdgit clone --depth 1 https://github.com/glslang/windbg-mcpWrote 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/glslang/windbg-mcp/eval-bench)<a href="https://agentmods.dev/skills/glslang/windbg-mcp/eval-bench"><img src="https://agentmods.dev/badge/skills/glslang/windbg-mcp/eval-bench.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 | $0.00084 | $0.04192 |
| Opus 5 | $0.00042 | $0.02096 |
| Sonnet 5 | $0.00017 | $0.00838 |
| Haiku 4.5 | $0.00008 | $0.00419 |
Grade A, and why
eval-bench 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 today.
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 — 207 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Benchmarking a model against this server
The grid, and what bites while running it
docs/local-model-eval.md is the result; this is what bites while running it again. The grid is
three scripts — the ollama driver, the Claude Code driver, and the matrix runner that spawns either
one per cell and grades the log afterwards.
Record what the runtime served, not what you asked for. num_ctx on a request does not
shrink an instance ollama already holds: with a 32,768 instance loaded, cells asking for 8,192 are
served 32,768 and look perfectly healthy — a 17,300-token prompt "fitting" in 8k, which is the
result the context axis exists to find and would have been fiction. /api/ps is the only place
the truth appears. Every record carries served_context, the grader marks a cell where the two
disagree with ?, and the runner evicts the model between windows. The first run of the grid
recorded five such cells; they were dropped and re-run.
The grader's three matching rules each came from a real wrong verdict, and all three are in
present():
- A number matches only between hex boundaries —
0x22is the device typeioctl_decodeasks for and0x22200Bis the code in the question, so plain containment passed for any answer that repeated the question. One model scored correct while sayingFILE_DEVICE_KEYBOARD. - Leading zeros are formatting — the tool prints
0x802and a model writing0x0802agrees with it. That one marked a correct control answer wrong. - A separator between hex digits is formatting too. WinDbg writes
fffff801`3c65bca8; Opus writes0xfffff801_3c65bca8. Both name the address the key holds.
Two of those three were found by reading the control's answers, which is the argument for having a frontier row at all.
possible_on in the task file is a prediction, and predictions about this server are wrong in
one direction: too pessimistic. Facts here are reachable by more than one route —
open_dump's summary carries the bug check and the module count, crash_triage's frame 0 is the
pc that registers reports — so a task the tool table says needs inspect may be answerable
with crash alone. Verify against the dump before scoring a model wrong for finding the other
route; the arm64_pc entry was corrected mid-run for exactly this.
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.
- today First seen · 207 lines · 84 tokens per session scan A 77b845582bd3
eval-bench is a skill published in the GitHub repository glslang/windbg-mcp (9 stars, last pushed yesterday), licensed MIT. It adds 84 tokens to every session and 4,192 once invoked, about $0.0004 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 2026-09-06.
Other skills, from other repositories
agent-platform-rag-engine-management
Manage and query Agent Platform RAG Engine Corpora and retrieve grounded contexts using the Google GenAI SDK. Use when listing RAG corpora or files, inspecting a corpus, retrieving contexts, or generating content grounded in a RAG corpus. Do not use for standard database queries (use SQL/Spanner skills), Google…
agent-platform-model-registry
Agent Platform Model Registry Management. Use when you need to upload, list, describe, update, or delete machine learning models (and their versions) in the Agent Platform Model Registry. Don't use for model training, model deployment to endpoints, or managing non-Agent Platform models.
foundry-config-setup
Resolve missing setup caused by a hardcoded Foundry project endpoint or model in a sample. Use when a sample fails because it uses a placeholder/hardcoded projectendpoint (for example "https://your-project.services.ai.azure.com") or a hardcoded model instead of reading them from the environment.
google-cloud-solution-agentic-analytics-spark-knowledge-catalog
Discovers requirements and generates guidance to design and deploy a governed, secure agentic-analytics solution for data that's distributed across Google Cloud, other cloud providers, or on-premises. Data that's outside Google Cloud (such as data from Databricks, Snowflake, Salesforce, SAP, or Oracle systems) is…
training-check
Interactively monitor training metrics from the current Codex session, periodically checking WandB or fallback logs for NaN, divergence, plateaus, and broken runs.
nemo-automodel-launcher-config
Configure NeMo AutoModel job launches for interactive runs, Slurm clusters, and SkyPilot cloud execution.