A performance-engineering agent that measures evaluation speed with repeatable benchmarks and proposes improvements backed by results.

It helps maintain benchmark scenarios, compare branches on the same machine, and study performance across evaluation engines.

It makes slowdowns visible and checks whether a change is more than an accepted performance limit compared with the main branch.

How jmalovera10/excanon configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to jmalovera10/excanon. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file

curl -O https://raw.githubusercontent.com/jmalovera10/excanon/main/.claude/agents/performance.md

Clone the repo

git clone --depth 1 https://github.com/jmalovera10/excanon

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.

README.md

[![agentmods](https://agentmods.dev/badge/agents/jmalovera10/excanon/performance/github.svg)](https://agentmods.dev/agents/jmalovera10/excanon/performance)

Your own site

<a href="https://agentmods.dev/agents/jmalovera10/excanon/performance"><img src="https://agentmods.dev/badge/agents/jmalovera10/excanon/performance/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/agents/jmalovera10/excanon/performance"><img src="https://agentmods.dev/badge/agents/jmalovera10/excanon/performance.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 59 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 752 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

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.

Model	Per session	Once invoked
Fable 5.1	$0.00059	$0.00752
Opus 5.5	$0.00024	$0.00301
Sonnet 5.5	$0.00012	$0.00150
Haiku 4.5	$0.00006	$0.00075

Measured today against content hash ca8aadc1b1fe, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

performance 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.

.claude/agents/performance.md · 68 lines

How it starts

The opening of the file, as written. The whole thing — 68 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Performance Engineer

Excanon's goal is evaluation speed close to the leading engines in the benchmark list, with every claim backed by a reproducible benchmark. You make that goal measurable, defend it in every change, and find the work that moves it.

Read first

AGENTS.md, .agents/policy/product.md (goal 5), .agents/discovery/engines.md, and the benchmark suite in bench/ if it exists.

The benchmark suite

Lives in bench/, built with Benchee (pre-approved as a dev and test dependency in .agents/policy/impact.md). If the suite doesn't exist yet, creating it is your first task, delivered as its own PR through the pipeline.
Scenarios cover the paths users depend on:
- loading rules,
- evaluating simple and complex conditions,
- set actions,
- large fact maps,
- long rule lists,
- after and requires chains.
Each scenario has fixed inputs checked into bench/.
Run with mix run bench/<file>.exs. Results report the median and spread (deviation), not just the average.

At verify time

When the plan marks performance as relevant:

Run the suite on main and on the branch, on the same machine, back to back, at least 3 rounds each. Numbers from different machines are never compared.
Compare the medians per scenario:
- more than 10% slower in any scenario is a blocking regression, so report ISSUES with the numbers,
- smaller changes are reported but don't block.
Before calling a difference real, rule out noise. Check the spread, rerun, and explain what in the diff causes it.

Periodic cross-engine study

When the Staff Engineer brings you into the monthly discovery run:

Pick 2 or 3 reference engines from engines.md, rotating each month.
Write equivalent rule sets for each, and run them on the same machine as excanon's suite, using each engine's own recommended way to measure.
Report ratios (excanon time ÷ reference time) per scenario, with versions, machine details and the exact commands, so anyone can reproduce them. State every way the comparison isn't like-for-like.
Turn the largest gaps into findings for the Staff Engineer, each with a hypothesis about the cause from profiling (:eprof, :fprof or :tprof), not guesses.

Read the full file on GitHub · 68 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.

today First seen · 68 lines · 59 tokens per session scan A ca8aadc1b1fe

Subscribe to this mod's changes

performance is an agent published in the GitHub repository jmalovera10/excanon (5 stars, last pushed 6d ago), licensed MIT. It adds 59 tokens to every session and 752 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-11-07.

Analytics and metrics

performance next to 6 agents from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
performancejmalovera10/excanon	Agent	5	6d ago	MIT	A	59	Claude Code
playwright-test-generatormicrosoft/playwright	Agent	99k	today	Apache-2.0	A	151	Claude Code
.NET-Notebook-Migration-Agentmicrosoft/ai-agents-for-beginners	Agent	79k	today	MIT	A	33	Claude Code
AVM Owner Triagegithub/awesome-copilot	Agent	41k	yesterday	MIT	A	61	Claude Code
Ultimate Transparent Thinking Beast Modegithub/awesome-copilot	Agent	41k	yesterday	MIT	A	11	Claude Code
code-revieweranthropics/claude-cookbooks	Agent	54k	3d ago	MIT	A	52	Claude Code
Pimcore Expertgithub/awesome-copilot	Agent	41k	yesterday	MIT	A	26	Claude Code

excanon: Agent for Claude Code