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 skills add OctoPerf/octoperf-claude-plugins --skill octoperf-onpremise-agentgit clone --depth 1 https://github.com/OctoPerf/octoperf-claude-pluginsWrote 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.
[](https://agentmods.dev/skills/octoperf/octoperf-claude-plugins/octoperf-onpremise-agent)<a href="https://agentmods.dev/skills/octoperf/octoperf-claude-plugins/octoperf-onpremise-agent"><img src="https://agentmods.dev/badge/skills/octoperf/octoperf-claude-plugins/octoperf-onpremise-agent/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.
<a href="https://agentmods.dev/skills/octoperf/octoperf-claude-plugins/octoperf-onpremise-agent"><img src="https://agentmods.dev/badge/skills/octoperf/octoperf-claude-plugins/octoperf-onpremise-agent.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00000 | $0.01896 |
| Opus 5 | $0.00000 | $0.00948 |
| Sonnet 5 | $0.00000 | $0.00379 |
| Haiku 4.5 | $0.00000 | $0.00190 |
Grade D, and why
octoperf-onpremise-agent scanned grade D with 3 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 12d 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.
Asks for rootmediumPrivilege escalation
A mod that escalates privileges can change anything on the machine, not only the project.
(`sudo docker rm -f <name>`) on its machine, then `get_onpremise_agent_command` Downloads and executes remote codehighSupply chain
curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.
installed, tell them to run `wget -qO- https://get.docker.com/ | sh` first. Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
installed, tell them to run `wget -qO- https://get.docker.com/ | sh` first. How it starts
The opening of the file, as written. The whole thing — 139 lines — stays where its author put it; the contents beside it link to each section on GitHub.
On-premise providers & agents — capacity-first playbook
Set up and run OctoPerf load tests from your own machines instead of the Cloud. This is the deeper playbook behind the on-premise tools; read it before creating a provider or installing an agent.
Mental model — a provider is a localized load capacity
Think in virtual users (VUs) per location — from machines that can reach the target — not megabytes.
- A provider supplies load generators through agents you install on your own machines. It does not itself run load generators — it is the configuration (memory sizing) plus the set of locations the agents attach to.
- Each location is a named place (a city, region or country) with map coordinates. A test can spread load across locations (e.g. Paris + Madrid).
- An agent is one Docker container you run on one machine. At run time an agent starts load generators on demand, several per machine as memory allows — so one agent handles many VUs, not one.
- The provider's memory config is inherited by every load generator, so
RAM per machine × usable % ÷ memory-per-VU ≈ VUs per agent. Total provider capacity = that, times the number of agents, per location.
Capacity is what the user cares about ("I want 10 000 VUs in Paris, 2 000 in Madrid"). The memory config is just the means.
Accessibility — agents must reach the target (and stay off it)
On-premise exists mostly for reach: the OctoPerf Cloud cannot hit an app that is internal, behind a firewall or on a private network. The fix is to run agents on machines that sit inside that network.
- Every agent machine must be able to reach the application under test — same LAN / VPN / VPC, DNS resolves, ports open. An agent that can't reach the target only produces connection errors. This reachability, more than geography, is usually the real reason a location exists.
- Never install an agent on the machine that hosts the app under test. The agent spawns load generators there; their CPU/RAM steal from the app server and skew every result. Use a separate machine on the same network.
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.
- 12d ago First seen · 139 lines · 0 tokens per session scan D 63ecd8663704
octoperf-onpremise-agent is a skill published in the GitHub repository OctoPerf/octoperf-claude-plugins (0 stars, last pushed 4d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 1,896 tokens. A static security scan graded it D with 3 findings (asks for root, downloads and executes remote code, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
forge-app-review
Performs a lightweight pre-release readiness review of Atlassian Forge apps across manifest/module wiring, architecture, runtime compatibility, dependency posture, tests, deploy readiness, and obvious security, cost, or reliability smells. Use when the user asks "review my Forge app", "pre-deploy check", "is this app…
run-iteration-eval
Run the Logic-Lens content-eval pipeline for one iteration and produce a scored summary.json — use to measure a skill change. Wraps scripts/run-content-evals.sh (runner, costs tokens) and scripts/grade-iteration.py (grader, free, re-runnable). ALWAYS sync the plugin cache first. Use when the user wants to "run the…
spec-driven-dev
Use when starting any non-trivial feature, refactor, or new project that will touch more than one file. Drives an AI coding agent through a gated Spec → Plan → Build → Test → Review → Ship lifecycle so work is specified before it is built, verified before it is reviewed, and reviewed before it ships.
iterate-skill
Run the Logic-Lens skill-improvement loop end to end — baseline → diagnose failures → edit → sync cache → re-eval → verify net gain → iterate until clean. Use whenever the goal is to RAISE a skill's eval score or fix a failing eval mode: "improve logic-review", "the format compliance is failing, fix it", "iterate on…
sumup-testing
Set up and run SumUp sandbox tests. Use when configuring a SumUp test merchant, picking test cards, triggering deliberate failure (amount = 11), or building a SumUp end-to-end test harness.
python-testing
Python testing strategies using pytest, TDD methodology, fixtures, mocking, parametrization, and coverage requirements.