grader

grader is an agent for Claude Code from gokeshenzhen/gen-tb-skill. It costs 0 tokens per session (1,491 once invoked), scanned A, original, Apache-2.0.

Instructions for an evaluation grader that assesses generated hardware verification testbenches after automated checks have already run. It judges qualities such as UVM structure and semantic fidelity rather than repeating compile or simulation assertions.

In plain words
What is it for?
Use it to inspect evaluation outputs, transcripts, assertion results, and optional reference files, then rate the testbench across defined quality dimensions.
Why use it?
It separates machine-checkable results from expert review of design quality and correctness.

Agent for Claude Code

Written for Claude Code: a Claude Code subagent (agents/*.md). Also seen: mentions CLAUDE.md; mentions subagents.

Install

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.

agentmods
npx agentmods add agents/gokeshenzhen/gen-tb-skill/grader
Clone the repo
git clone --depth 1 https://github.com/gokeshenzhen/gen-tb-skill

Made for: Claude Code.

Wrote 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.

agentmods badge for grader

README.md
[![agentmods](https://agentmods.dev/badge/agents/gokeshenzhen/gen-tb-skill/grader.svg)](https://agentmods.dev/agents/gokeshenzhen/gen-tb-skill/grader)
Your own site
<a href="https://agentmods.dev/agents/gokeshenzhen/gen-tb-skill/grader"><img src="https://agentmods.dev/badge/agents/gokeshenzhen/gen-tb-skill/grader.svg" alt="Measured on agentmods" height="20"></a>
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,491 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00000 $0.01491
Opus 5 $0.00000 $0.00745
Sonnet 5 $0.00000 $0.00298
Haiku 4.5 $0.00000 $0.00149

Measured 6d ago against content hash cd86a04f864a, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

grader 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 6d 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.

evals/agents/grader.md · 117 lines

How it starts

The opening of the file, as written. The whole thing — 117 lines — stays where its author put it; the contents beside it link to each section on GitHub.

gen-tb Grader Agent

You are the quality grader for a gen-tb eval run. The mechanical assertions (compile_exit_zero, sim_passes, log_contains, …) have already been judged by scripts/run_evals.py and saved to assertions_result.json. Do not re-judge them. Your job is to evaluate dimensions those assertions can't catch.

Inputs (paths are given to you in the user prompt)

  • outputs_dir: directory with the generated UVM tb (tb/, test/, top/, script/, work/, CLAUDE.md, .prj_top)
  • transcript_path: transcript.md with the run's prompt, scaffold log, compile log, sim log tails
  • assertions_result_path: machine-judged assertion verdicts (read, don't re-judge)
  • expected_dir (optional, may be missing): canonical reference under evals/fixtures/<name>/expected/ — use to spot-check semantic fidelity, not to do a diff-against-expected. The skill is not required to reproduce expected/ byte-for-byte.

Quality dimensions to evaluate

For each dimension below, decide a verdict in {strong, ok, weak, broken} with cited evidence:

Dimension What to look for
uvm_style sequencer/sequence/driver/monitor/agent/env/scoreboard roles cleanly separated; factory registration consistent; config_db paths sensible; phase usage correct
scoreboard_value scoreboard actually compares DUT vs ref/expectation. Red flag: empty check_phase, always-pass logic, or no reference at all — that would let sim_passes succeed vacuously
ral_correctness RAL is 1:1 with work/_gen_audit/spec_normalized/registers.yaml (no invented or dropped registers); access policies match; aliased registers / arrays handled
tb_api_bfm tb_api::write/read/expect_reg task-style BFM exists and is usable by DE persona — not just a thin wrapper that requires a UVM context
hardcoding_risk tests achieve log_contains needles by genuine reads/checks, not by $display-ing the magic string. Look for the needle in the test source — if it's a literal in a display/info, that's a cheat
unresolved_honesty work/_gen_audit/unresolved.md exists if there were genuine ambiguities; the skill didn't hide ambiguity by hardcoding a guess
directory_hygiene layout matches references/directory_layout.md; no stray files under fixture inputs; no symlinks pointing outside the IP root
generated_claude_md CLAUDE.md exists and tells a future maintainer the local conventions (file ownership, makefile API, how to add a test). A missing or trivial CLAUDE.md is a weak.
generic_mode_honesty Only when bus_protocol: generic in work/_gen_audit/scaffold_audit.json. Does the tb log every protocol assumption the sub-agent made? work/_gen_audit/generic_bus_scaffold_prompt.md must contain a populated ## Assumptions made by sub-agent section (not just the empty heading scaffold.py writes). The generated CLAUDE.md must carry the Generic-mode review checklist verbatim from references/generic_bus.md. Red flag: tb compiles + sim passes but no assumption list — that means the sub-agent silently picked interpretations a reviewer cannot audit. On built-in buses, report ok with evidence "not applicable: bus_protocol is <bus>".

Read the full file on GitHub · 117 lines

Changes

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.

  1. 6d ago First seen · 117 lines · 0 tokens per session scan A cd86a04f864a

Subscribe to this mod's changes

grader is an agent published in the GitHub repository gokeshenzhen/gen-tb-skill (23 stars, last pushed 1mo ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 1,491 tokens. 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.

Related

Other agents, from other repositories

apple-neural-performance-expert

Use this agent when you need expert guidance on optimizing neural network operations on Apple platforms, including Metal Performance Shaders (MPS), MLX framework optimization, low-level array operations, GPU kernel optimization, memory management for ML workloads, or performance profiling of neural network code. This…

FluidInference/FluidAudio · 0 tokens

antenna-engineer

Reasons from gain–directivity–efficiency, Chu–Harrington bandwidth limits, and array factor through HFSS/CST/FEKO synthesis, IEEE 149-2021 NF/FF/CATR metrology, CTIA TRP/TIS/ECC OTA, and Friis link budgets while treating ground-plane truncation, active impedance in arrays, range ripple, and S₁₁≠pattern conflation as…

K-Dense-AI/scientific-agents · 97 tokens

fabricator

3D printing and additive manufacturing specialist covering FDM, SLA, and SLS processes from model preparation through troubleshooting.

pjt222/agent-almanac · 25 tokens

cocotb-reviewer

RAT audit protocol (condensed; dev source: plugindocs/agent-lib/audit-output-protocol.md — plugin-internal, do NOT Read it at runtime).

babyworm/rtl-agent-team · 51 tokens

labview-class-generator

Creates LabVIEW classes end to end — settles the data model, writes each .lvclass with its private data control through NI's own project provider VIs, links inheritance, creates INTERFACES and links a class to the ones it implements, binds .ctl typedef fields so they point at the file rather than carrying a de-linked…

Zuehlke/labview-mcp · 404 tokens

persona-high

Simulated senior IC designer with full datasheet / PDK / corner fluency. Specifies CRC polynomials, bit-period cycles, opcode hex, GF180MCU 5V corners. Pushes back hard when the AI hand-waves and demands datasheet-section traceability. Drives the IC Expert Agent (plain-language register) during Phase-1 training to…

vibeic/vibe-ic · 111 tokens