eval-prompts

A command that runs a prompt-evaluation test suite for a generative-AI application and reports whether its outputs meet release thresholds.

In plain words
What is it for?
Useful for scoring test cases, reviewing results by category, and identifying the cases that caused a failed release check.
Why use it?
It exposes problems with factual grounding, usefulness, clarity, or safety before the application ships.

Command for Claude Code

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 commands/timothywarner-org/claude-code/eval-prompts
Clone the repo
git clone --depth 1 https://github.com/timothywarner-org/claude-code

Made for: Claude Code.

Per session 14 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 235 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 $0.00014 $0.00235
Opus 5 $0.00007 $0.00118
Sonnet 5 $0.00003 $0.00047
Haiku 4.5 $0.00001 $0.00023

Measured 3d ago against content hash 49777b0a07dc, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

eval-prompts 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 3d 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.

.claude/commands/eval-prompts.md · 14 lines

What it actually says

/eval-prompts

Score the current GenAI app's outputs before it ships. Eval cases file: $1 (default to the genai-prompt-eval skill's resources/templates/eval_cases.jsonl if empty).

  1. Invoke the genai-prompt-eval skill.
  2. Load the eval cases from $1, run each through the model, and score the four dimensions: groundedness, relevance, coherence, safety.
  3. Report a per-dimension score, the pass threshold, and a clear PASS or FAIL per dimension. Do not signal pass/fail by color alone; use the words.
  4. On any FAIL, name the specific cases that dragged the score down so the prompt or grounding can be fixed. This is the same gate /deploy-genai runs before it ships. See [[testing]] and [[genai-prompt-eval]].
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. 3d ago First seen · 14 lines · 14 tokens per session scan A 49777b0a07dc

Subscribe to this mod's changes

eval-prompts is a command published in the GitHub repository timothywarner-org/claude-code (223 stars, last pushed 1mo ago), licensed MIT. It adds 14 tokens to every session and 235 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 2026-08-30.