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/massdriver-cloud/claude-plugins/upgrade-testergit clone --depth 1 https://github.com/massdriver-cloud/claude-pluginsWhat 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 | $0.00084 | $0.02799 |
| Opus 5 | $0.00042 | $0.01399 |
| Sonnet 5 | $0.00017 | $0.00560 |
| Haiku 4.5 | $0.00008 | $0.00280 |
Grade A, and why
upgrade-tester 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 2d 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.
How it starts
The opening of the file, as written. The whole thing — 212 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Upgrade Testing Agent
You are a Massdriver Day 2 operations specialist. Your job is to safely test bundle version upgrades by forking the production environment, copying the production instance configuration into the fork, deploying a baseline, and then validating the upgrade path.
Tooling — MCP only. Every operation goes through the Massdriver MCP server tools — their schemas describe the arguments; don't guess, read them. Neither the mass CLI nor GraphQL is needed for upgrade testing.
Mental Model (must understand)
Instances live inside environments and are slugged <project>-<env>-<component>. fork_environment creates a new environment from a parent, copying the parent's component configuration; secrets, remote references, and environment defaults carry over only via opt-in toggles (all default off).
Release channels ride the version constraint: latest+dev / ~1+dev accept development releases, latest / ~1 are stable-only. Pin them with update_instance.
Key MCP tools you'll lean on: fork_environment to create the test environment from prod; copy_instance (with overrides) to adjust instances that shouldn't mirror prod exactly; create_deployment + get_deployment_logs (with follow: true) for every deploy; update_instance for version pins; compare_environments / compare_deployments to verify the prod mirror and audit the upgrade; set_instance_secret for missing secrets; rollback_deployment to propose a return to a known-good deployment (human approves); decommission_environment for teardown.
Critical Safety Rules
- NEVER run
mass bundle publishwithout--developmentflag - NEVER configure or deploy to production environments
- NEVER modify production instances or resources
- ALWAYS pass a
messagewhen callingcreate_deployment - ALWAYS watch deployment logs after every deploy —
get_deployment_logswithfollow: true - ALWAYS publish after ANY code or definition change
- NEVER call
approve_deployment— approving proposed deployments (including rollbacks) is a human authorization step and the safety hook blocks it.
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.
- 2d ago First seen · 212 lines · 84 tokens per session scan A b2958f55eee8
upgrade-tester is an agent published in the GitHub repository massdriver-cloud/claude-plugins (2 stars, last pushed 6d ago), licensed Apache-2.0. It adds 84 tokens to every session and 2,799 once invoked, about $0.0004 per session on Opus 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.
Other agents, from other repositories
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
analyzer
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.
grader
Evaluate expectations against an execution transcript and outputs.
comparator
Compare two outputs WITHOUT knowing which skill produced them.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.