benchmarks: Skill for Claude Code

.agents/skills/local-platform-e2e/SKILL.md

local-platform-e2e is a skill for Claude Code, Codex from computesdk/benchmarks. It costs 82 tokens per session (3,692 once invoked), scanned A, original, MIT.

A procedure for running benchmark tests against a local copy of the benchmarks platform. The setup uses local database and object-storage services, so it does not require cloud access or provider credentials.

In plain words
What is it for?
Use it to start the platform locally and run a real `@benchsdk/runner` benchmark. It covers creating a benchmark, running workers, recording results, uploading artifacts, and completing the run.
Why use it?
It lets developers test the full benchmark workflow without depending on external services. It helps investigate problems in worker planning, task reporting, artifact uploads, and run completion.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: installed under .agents/ (shared by several agents).

This is computesdk/benchmarks's own configuration. It tells Claude Code and Codex how to work on benchmarks itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything benchmarks configures →

Reuse

Borrowing it

Nothing to install: this file belongs to computesdk/benchmarks. 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/computesdk/benchmarks/master/.agents/skills/local-platform-e2e/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/computesdk/benchmarks

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.

agentmods badge for local-platform-e2e

README.md
[![agentmods](https://agentmods.dev/badge/skills/computesdk/benchmarks/local-platform-e2e/github.svg)](https://agentmods.dev/skills/computesdk/benchmarks/local-platform-e2e)
Your own site
<a href="https://agentmods.dev/skills/computesdk/benchmarks/local-platform-e2e"><img src="https://agentmods.dev/badge/skills/computesdk/benchmarks/local-platform-e2e/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.

agentmods 80×15 button for local-platform-e2e

Your own site · 80×15
<a href="https://agentmods.dev/skills/computesdk/benchmarks/local-platform-e2e"><img src="https://agentmods.dev/badge/skills/computesdk/benchmarks/local-platform-e2e.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 82 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 3,692 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. A grade says what 26 rules found in the file — not that it is safe. Third-party audits
  • NVIDIA SkillSpector warn 7 Sept 2026
SkillSpector: 9 findings, up to high

These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →

  • high Privilege Escalation · line 19
    Potential security issue detected. Manual review is recommended.
    Fix: Review the flagged content for security risks. Ensure no credentials, secrets, or sensitive data are exposed.
  • high Privilege Escalation · line 97
    Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
    Fix: Remove references to credential paths. Use environment variables or secrets managers. For docs, use placeholder paths (e.g., /path/to/config). Never load .env or token files in production code paths.
  • medium MCP Rug Pull · line 15
    Docker image references without a specific tag (:latest is implicit) or digest (@sha256:...) can be silently replaced by a malicious image.
    Fix: Pin the image: image:tag or image@sha256:abc123
  • medium MCP Rug Pull · line 16
    Docker image references without a specific tag (:latest is implicit) or digest (@sha256:...) can be silently replaced by a malicious image.
    Fix: Pin the image: image:tag or image@sha256:abc123
  • medium MCP Rug Pull · line 18
    Docker image references without a specific tag (:latest is implicit) or digest (@sha256:...) can be silently replaced by a malicious image.
    Fix: Pin the image: image:tag or image@sha256:abc123
  • medium MCP Rug Pull · line 19
    Docker image references without a specific tag (:latest is implicit) or digest (@sha256:...) can be silently replaced by a malicious image.
    Fix: Pin the image: image:tag or image@sha256:abc123
  • medium Data Exfiltration · line 34
    Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
    Fix: Verify the destination URL is trusted and necessary. Remove or replace with documented APIs. Ensure no secrets, tokens, or PII are transmitted.
  • medium MCP Rug Pull · line 97
    npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.
    Fix: Pin the version: npx @scope/[email protected]
  • medium MCP Rug Pull · line 150
    npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.
    Fix: Pin the version: npx @scope/[email protected]
How audits are shown
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 5.1 $0.00082 $0.03692
Opus 5.5 $0.00033 $0.01477
Sonnet 5.5 $0.00016 $0.00738
Haiku 4.5 $0.00008 $0.00369

Measured 6d ago against content hash 54b0586e84b0, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

local-platform-e2e scanned grade A with 1 finding 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 6d 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.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

curl -s "http://127.0.0.1:8123/?user=default&password=chpass" \
.agents/skills/local-platform-e2e/SKILL.md · 246 lines

How it starts

The opening of the file, as written. The whole thing — 246 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Local end-to-end: @benchsdk/runner ↔ benchmarks-platform

Goal: exercise upsert benchmark → create run → planWorkers → claim → heartbeat → task_results → artifact upload → complete → dashboard, with zero external credentials.

1. Infra (docker)

docker run -d --name pg -p 5432:5432 -e POSTGRES_PASSWORD=postgres -e POSTGRES_DB=bench postgres:16
docker run -d --name minio -p 9000:9000 -e MINIO_ROOT_USER=minioadmin -e MINIO_ROOT_PASSWORD=minioadmin \
  quay.io/minio/minio server /data
docker run -d --name ch -p 8123:8123 -e CLICKHOUSE_PASSWORD=chpass clickhouse/clickhouse-server:25.6
docker run --rm --network host --entrypoint sh quay.io/minio/mc -c \
  "mc alias set local http://127.0.0.1:9000 minioadmin minioadmin && mc mb -p local/bench"

Substitutes when images or ports aren't available:

  • Postgres port taken (a host postgres already on 5432): use it — DATABASE_URL=postgresql://ci:[email protected]:5432/ci (the CI creds; the ci database works when CREATE DATABASE is denied).

  • MinIO image won't pull (pull access denied / quay 401): s3rver is a drop-in — npx -y s3rver -p 9000 -a 0.0.0.0 --directory ~/s3data, then AWS_ACCESS_KEY_ID=S3RVER AWS_SECRET_ACCESS_KEY=S3RVER aws --endpoint-url http://127.0.0.1:9000 s3 mb s3://bench and set TIGRIS_ACCESS_KEY_ID/ TIGRIS_SECRET_ACCESS_KEY=S3RVER in .env.local.

  • npm run db:migrate says DATABASE_URL ... is not set: it doesn't read .env.local — run with set -a; . ./.env.local; set +a; npm run db:migrate.

  • Minting /mcp access tokens for MCP E2E: insert a jwks row (EdDSA keypair, both JWKs stored), an oauth_client row, and an oauth_consent row (clientId + userId + referenceId=orgId, non-null scopes), then sign an at+jwt (kid = the jwks row id) with claims sub=user, client_id, org, scope, iss=$BETTER_AUTH_URL/api/auth, aud=$BETTER_AUTH_URL/mcp — see lib/oauth/tokens.ts verifyAccessToken for the exact shape.

  • Use ClickHouse >= 25.x: 24.8 fails ch:migrate with TTL expression result column should have DateTime or Date type, but has DateTime64(3,'UTC').

  • The importer passes clickhouse_settings: { date_time_input_format: 'best_effort' } per insert, so a default-configured server works. If you hit Cannot parse input: expected '"' before: 'Z"...' (older code), work around it server-side, and remember to REMOVE the override before verifying an importer fix — otherwise the server setting masks it:

    docker exec ch bash -c 'mkdir -p /etc/clickhouse-server/users.d && printf "<clickhouse><profiles><default><date_time_input_format>best_effort</date_time_input_format></default></profiles></clickhouse>" > /etc/clickhouse-server/users.d/besteffort.xml'
    docker restart ch
    # verify which mode is actually active:
    curl -s "http://127.0.0.1:8123/?user=default&password=chpass" \
      --data-binary "SELECT value FROM system.settings WHERE name='date_time_input_format'"
    

Read the full file on GitHub · 246 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.

  1. 6d ago Changed · +19 lines 54b0586e84b0
  2. 2mo ago First seen · 227 lines · 82 tokens per session scan A 78b4c0a322c7

Subscribe to this mod's changes

local-platform-e2e is a skill published in the GitHub repository computesdk/benchmarks (131 stars, last pushed yesterday), licensed MIT. It adds 82 tokens to every session and 3,692 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.