What is copilot-benchmark?

A benchmark suite for testing how well the Copilot command-line agent answers questions about a code repository, with and without Lore tools.

It is for running repeated benchmark tasks against repositories such as Zod, FastAPI, esbuild, Postgres, and Gson, then comparing the results.

It measures whether the added tools improve answer correctness, coverage, and efficiency rather than relying on impressions.

How jafreck/Lore configures Claude Code and Codex

Reuse

Borrowing it

Nothing to install: this file belongs to jafreck/Lore. 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/jafreck/Lore/main/.github/skills/copilot-benchmark/SKILL.md

Clone the repo

git clone --depth 1 https://github.com/jafreck/Lore

Made for: Claude Code, Codex.

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/skills/jafreck/lore/copilot-benchmark/github.svg)](https://agentmods.dev/skills/jafreck/lore/copilot-benchmark)

Your own site

<a href="https://agentmods.dev/skills/jafreck/lore/copilot-benchmark"><img src="https://agentmods.dev/badge/skills/jafreck/lore/copilot-benchmark/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/skills/jafreck/lore/copilot-benchmark"><img src="https://agentmods.dev/badge/skills/jafreck/lore/copilot-benchmark.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 43 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 1,655 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	$0.00043	$0.01655
Opus 5.5	$0.00017	$0.00662
Sonnet 5.5	$0.00009	$0.00331
Haiku 4.5	$0.00004	$0.00166

Measured 1mo ago against content hash 5c9003ad7087, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

copilot-benchmark 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 1mo 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.

.github/skills/copilot-benchmark/SKILL.md · 129 lines

How it starts

The opening of the file, as written. The whole thing — 129 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Copilot Agent Benchmark

Purpose

Run Lore's Copilot agent benchmark harness, which evaluates how the Copilot CLI answers codebase questions with and without Lore MCP tools, comparing the two arms on correctness, coverage, and efficiency.

Prerequisites

copilot CLI installed and authenticated (copilot --version must work).
Node.js 22 (use nvm use 22).
Lore built (npm run build) — the test beforeAll also runs this.
Real API calls are made — this costs tokens.

Quick start

source ~/.nvm/nvm.sh && nvm use 22
npm run build
BENCHMARK_COPILOT=1 npx vitest run tests/benchmark/copilot-agent.test.ts

Environment variables

Variable	Default	Description
`BENCHMARK_COPILOT`	(unset)	Required. Set to `1` to enable the suite (skipped otherwise).
`BENCHMARK_REPO`	`lore-self`	Target repo. Options: `lore-self`, `zod`, `fastapi`, `esbuild`, `postgres`, `jackson-databind`.
`BENCHMARK_MODEL`	`claude-opus-4.6`	LLM model passed to copilot CLI `--model`.
`BENCHMARK_INDEX_MODE`	`scip`	Index mode label: `scip` enables SCIP, `full` enables SCIP+LSP, and legacy `tree-sitter` currently produces file snapshots only because tree-sitter has been removed.
`BENCHMARK_ITERATIONS`	`1`	Runs per task. Use `≥3` for statistical significance.
`BENCHMARK_EMBEDDING_MODEL`	(empty)	Embedding model, e.g. `nomic-ai/nomic-embed-text-v1.5`.
`BENCHMARK_LSP`	(unset)	Set to `1` to enable LSP enrichment during indexing.
`BENCHMARK_QUESTION`	(empty)	Optional comma-separated question IDs to run.
`BENCHMARK_ARM`	(empty)	Optional single-arm filter: `control` or `lore`.

Instructions

When the user asks to run, execute, or launch a Copilot benchmark:

Pre-flight checks
- Ensure Node.js 22 is active: source ~/.nvm/nvm.sh && nvm use 22.
- Build Lore: npm run build.
- Verify copilot --version works.
Determine configuration from user request
- Pick a repo from the available list. Default is lore-self.
- Pick an index mode. Default to scip; do not use the legacy tree-sitter label when structural results are required.
- Pick iteration count. Default is 1 for quick runs, 3+ for statistical significance.
- Pick model. Default is claude-opus-4.6.

Read the full file on GitHub · 129 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.

1mo ago Changed · +3 lines 5c9003ad7087
2mo ago First seen · 126 lines · 43 tokens per session scan A 537c3fb4c705

Subscribe to this mod's changes

copilot-benchmark is a skill published in the GitHub repository jafreck/Lore (7 stars, last pushed 1mo ago), licensed MIT. It adds 43 tokens to every session and 1,655 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 2026-08-31.

Testing

research-engineer

An uncompromising Academic Research Engineer. Operates with absolute scientific rigor, objective criticism, and zero flair. Focuses on theoretical correctness, formal verification, and optimal implementation across any required technology.

davila7/claude-code-templates · 43 tokens

tika-eval-compare

Compare extracts from two Tika builds over a corpus to detect regressions in content, encoding, exceptions, and embedded-document handling. Use for "compare before/after extracts", "eval this change against the corpus".

apache/tika · 50 tokens

cua-driver

Use Cua Driver for desktop or browser tasks that are awkward or unavailable through Bash/APIs, or when the user explicitly wants GUI interaction: app testing, visual bug reproduction, form filling, calendar entry, screenshots, and demo recording. Also covers Cua setup; not OpenAI Codex Computer Use or web research.

davidondrej/skills · 68 tokens

Lore: Skill for Claude Code