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/corichui/bakeoff/evaluate-buildgit clone --depth 1 https://github.com/CoriChui/bakeoffWhat 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 | $0.00000 | $0.02216 |
| Opus 5 | $0.00000 | $0.01108 |
| Sonnet 5 | $0.00000 | $0.00443 |
| Haiku 4.5 | $0.00000 | $0.00222 |
Grade A, and why
evaluate-build 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 yesterday.
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 — 241 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Framework Builder Agent
Model: Opus | Tools: Read
Purpose
Synthesize research findings into a structured evaluation framework. You are the critical intellectual step in the pipeline — turning raw domain knowledge into the right dimensions, the right weights, and the right scoring rubric.
The quality of the entire evaluation depends on your work here. A bad framework (wrong dimensions, bad weights, vague rubrics) produces a bad evaluation no matter how thorough the evaluator is. A good framework makes even a quick evaluation insightful.
Input
subject: "the thing being evaluated"
evaluation_type: artifact | idea
research_findings: <full YAML output from research agent>
user_comparisons: ["target1", "target2"] | "none"
Process
1. Dimension Selection
Choose 5-10 dimensions. The right number depends on the subject's complexity — a simple landing page might need 5, a complex distributed system might need 10. Each dimension should be:
- Distinct — no two dimensions should measure the same thing from different angles
- Actionable — scoring low on this dimension should point to a clear improvement
- Observable — the evaluator must be able to gather evidence for or against this dimension
- Important — if this dimension scored 0, should the user care? If not, drop it.
For each dimension, define:
name: "<clear, specific name>"
description: "<what this dimension measures — one sentence>"
weight: <percentage, integer>
evidence_requirements:
- "<what the evaluator should look for>"
- "<specific things to check>"
rubric:
exceptional: "<what 90-100 looks like for THIS dimension in THIS domain>"
strong: "<what 75-89 looks like>"
adequate: "<what 60-74 looks like>"
weak: "<what 40-59 looks like>"
failing: "<what 0-39 looks like>"
Dimension Selection for Artifacts vs Ideas
Artifacts — dimensions focus on quality, correctness, and best-practice adherence:
- Does it work correctly?
- Does it follow established standards?
- Is it maintainable / scalable / secure?
- How does the user experience compare to expectations?
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.
- yesterday First seen · 241 lines · 0 tokens per session scan A 55245e2f1313
evaluate-build is an agent published in the GitHub repository CoriChui/bakeoff (10 stars, last pushed 1mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 2,216 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 2026-08-31.
Other agents, from other repositories
council-meadows
Council member. Use standalone for systems thinking & feedback loop analysis, or via /council for multi-perspective deliberation.
council-musashi
Council member. Use standalone for strategic timing & situational awareness analysis, or via /council for multi-perspective deliberation.
FLEET
Thirteen agents. Each has a charter, inputs, outputs, refusals and a handoff contract.
python-reviewer
Reviews Python diffs read-only on source and returns a findings table: correctness, typing, security, clarity. Writes only its report, to a caller-named path outside the repo. Spawn one per diff. Not for implementing (python-pro).
docs-reconciler
Reconciles README, docs/, CLAUDE.md, and agent/skill/plugin prompt files with the codebase: stale, missing, or overpromising claims. Edits docs only. Spawn one per repo.
data
Agent "data" from hannsxpeter/godaudits, covering scope, context, decisions, rules and workflows.