Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/genudo-ai/genudo_mcpnpx agentmods add skills/genudo-ai/genudo_mcp/run-agent-testsWrote 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/genudo-ai/genudo_mcp/run-agent-tests)<a href="https://agentmods.dev/skills/genudo-ai/genudo_mcp/run-agent-tests"><img src="https://agentmods.dev/badge/skills/genudo-ai/genudo_mcp/run-agent-tests/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/genudo-ai/genudo_mcp/run-agent-tests"><img src="https://agentmods.dev/badge/skills/genudo-ai/genudo_mcp/run-agent-tests.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 | $0.00088 | $0.00863 |
| Opus 5.5 | $0.00035 | $0.00345 |
| Sonnet 5.5 | $0.00018 | $0.00173 |
| Haiku 4.5 | $0.00009 | $0.00086 |
Grade A, and why
run-agent-tests 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 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.
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 — 76 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Run Agent Tests
Drive the engine, then explain what happened. Requires a working setup
(setup-testing-engine) and a suite (generate-test-journeys).
1. Confirm before any live run
A run creates real opportunities (flagged is_test), sends real messages to the agent, and
lets it fire real actions. State to the user: the pipeline(s), the case count, and which
webhooks may fire — then get a yes. Always do a dry run first:
python3 src/engine.py run --suite <suite.json> --pipeline <channel|all> --dry-run
2. Run
python3 src/engine.py run --suite <suite.json> --pipeline <channel|all> --run-no <N> \
[--cases J01,J02] [--priority P0|P0,P1|all] [--category TEXT] \
[--workers 3] [--stop-on-fail] [--keep-live] [--allow-prod]
--run-no— increment per run; it labels leads and typed namesTEST R<N>.- Targeted re-test after a fix:
--cases <failed ids>; regression:--priority P0. - Keep
--workersat 3 or less; the MCP limit is 30 requests/minute and the engine throttles and backs off on its own. --keep-liveskips cleanup — only when the user wants to inspect the test conversations in the inbox; remind them to archive afterwards.--allow-prodonly with explicit user approval for that run.
What the engine guarantees: each case gets its own @example.com test contact; it waits for
14s of agent silence (max 150s) before checking a step, because agents often answer in
several messages; and cleanup moves every test lead to Archived and pauses its conversation
even if the run crashes.
Exit code 1 means at least one P0 case did not pass.
3. Read the results
Output lands in runs/<run-id>/: results.json (every reply, check, action, stage, judge
verdict) and report.html. Summarize:
- Pass / fail / warn per priority, P0 first.
- Each failure: case and step, the customer message, what the agent did, which check failed (deterministic or judge), and the judge's reason.
- Steps marked
judge_erroror flagged only by Tier-1 checks → hand toevaluate-agent-repliesbefore calling them real defects.
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.
- 6d ago First seen · 76 lines · 88 tokens per session scan A 9f2d537db2e6
run-agent-tests is a skill published in the GitHub repository genudo-ai/genudo_mcp (0 stars, last pushed 6d ago), licensed MIT. It adds 88 tokens to every session and 863 once invoked, about $0.0004 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 2026-10-11.
Other skills, from other repositories
playwright-browser
Use when automating browsers, testing pages, taking screenshots, checking UI, verifying login flows, or testing responsive behavior.
local-e2e
A skill for running end-to-end tests against a local LeadAce application stack. End-to-end tests check complete user flows across the running application, rather than only testing individual pieces.
test-master
Generates test files, creates mocking strategies, analyzes code coverage, designs test architectures, and produces test plans and defect reports across functional, performance, and security testing disciplines. Use when writing unit tests, integration tests, or E2E tests; creating test strategies or automation…
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.
critique
Adversarial design critique of the current work — render it, look at it, and argue for rejection. Run after the gates are green, never instead of them.
test-warp-ui
Guides testing Warp UI features and changes using the computer use tool. Use this skill only when computer-use testing was requested (explicit request or accepted offer) and the computeruse tool is available to the agent. Covers launching Warp and verifying UI behavior.