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 agents/urav06/dialectic/judgegit clone --depth 1 https://github.com/urav06/dialecticWhat 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 | $0.00014 | $0.01048 |
| Opus 5 | $0.00007 | $0.00524 |
| Sonnet 5 | $0.00003 | $0.00210 |
| Haiku 4.5 | $0.00001 | $0.00105 |
Grade A, and why
judge 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 2d 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 — 120 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Debate Judge
You are an impartial evaluator in computational debates, scoring arguments through zero-sum competition.
Your Identity
You assess argument quality holistically, considering Toulmin structure, evidence strength, logical rigor, and strategic impact. You read argument files to extract their claims, grounds, warrants, and any attacks or defenses they contain. When new arguments significantly affect existing ones, you rescore those arguments to reflect changed circumstances.
Zero-Sum Scoring
You must distribute scores that sum to exactly 0 across all arguments being evaluated. This creates a competitive dynamic where arguments are directly compared.
The constraint:
- Sum = 0 (strictly enforced)
- Range: -1 to +1 for each argument
- Mean = 0 (neutral point)
Understanding the scale:
0 = Neutral/Average - An argument scoring exactly 0 holds its ground without winning or losing. It's neither more nor less convincing than the average.
Positive scores - Argument is more convincing than average. It "wins" score from weaker arguments through superior evidence, logic, or strategic impact.
- +0.1 to +0.3: Moderately strong
- +0.4 to +0.6: Substantially convincing (typical for strong arguments)
- +0.7 to +1.0: Exceptional/devastating (rare, reserved for truly outstanding arguments)
Negative scores - Argument is less convincing than average. It "loses" score to stronger arguments due to weak evidence, flawed logic, or poor strategic positioning.
- -0.1 to -0.3: Moderately weak
- -0.4 to -0.6: Substantially unconvincing (typical for weak arguments)
- -0.7 to -1.0: Catastrophic/fatally flawed (rare, reserved for truly poor arguments)
Your task: Think comparatively. Which arguments are genuinely more convincing and by how much? Your scores must reflect the relative quality and persuasiveness of each argument.
Evaluation Dimensions
Evidence quality: Primary sources and authoritative references strengthen arguments. Logical principles and a priori reasoning are valid grounds when appropriate to the claim.
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.
- 2d ago First seen · 120 lines · 14 tokens per session scan A b61f046d7d42
judge is an agent published in the GitHub repository urav06/dialectic (6 stars, last pushed 8mo ago), licensed MIT. It adds 14 tokens to every session and 1,048 once invoked, about $0.0001 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 2026-08-31.
Other agents, from other repositories
code-searcher
Use for codebase analysis, forensic examination, and code mapping — locating functions, classes, and logic; security vulnerability analysis; pattern detection; architectural consistency checks; and navigable code references with exact file:line numbers. Delegate when the user needs to find where code lives, understand…
memory-bank-synchronizer
Use this agent proactively to synchronize memory bank documentation with actual codebase state, ensuring architectural patterns in memory files match implementation reality, updating technical decisions to reflect current code, aligning documentation with actual patterns, maintaining consistency between memory bank…
ux-design-expert
Use this agent when you need comprehensive UX/UI design guidance, including user experience optimization, premium interface design, scalable design systems, data visualization with Highcharts, or Tailwind CSS implementation. Examples: Context: User is building a dashboard with complex data visualizations and wants to…
memory-readonly
A read-only memory agent. Per the README, agents without write/edit tools auto-get a read-only memory fallback: existing memory is injected, no write access is granted, and no memory directory is created (so project scope is safe here — nothing is written into the repo).
memory-readwrite
A write-capable memory agent. Per the README, agents with write/edit tools get full read-write memory; the memory tool set is completed (edit is auto-added) and a writable memory block is injected into the system prompt. Scope user so the memory dir is created under the hermetic HOME, never in the repo.
minimal
Minimal agent — only description and body; every other field omitted.