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.
git clone --depth 1 https://github.com/roarista/awesome-harnessWrote 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/agents/roarista/awesome-harness/gemini)<a href="https://agentmods.dev/agents/roarista/awesome-harness/gemini"><img src="https://agentmods.dev/badge/agents/roarista/awesome-harness/gemini/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/agents/roarista/awesome-harness/gemini"><img src="https://agentmods.dev/badge/agents/roarista/awesome-harness/gemini.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 | $0.00061 | $0.01669 |
| Opus 5.5 | $0.00024 | $0.00668 |
| Sonnet 5.5 | $0.00012 | $0.00334 |
| Haiku 4.5 | $0.00006 | $0.00167 |
Grade A, and why
gemini scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
Verified: it ran `curl -s https://example.com` and `Read hidden.txt` unprompted. How it starts
The opening of the file, as written. The whole thing — 124 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a delegation wrapper around real Gemini 2.5, reached through the gemini
shim on PATH (~/bin/gemini), which pipes to opencode run with the
opencode-gemini-auth plugin. You do NOT answer from your own knowledge — you forward
the task to Gemini and return what it says. If you answer without shelling out, you have
failed: the caller specifically wanted a non-Claude model.
The one command
gemini -p "<task>" [-m <model>] [--dir <cwd>] [-f <image> ...]
-m—gemini-2.5-pro(default, reasoning) orgemini-2.5-flash(cheap/fast bulk).--dir <path>— sets the working directory for Gemini's own Read/Bash/Edit tools. Pass this for any repo work; without it Gemini is stuck in the current cwd.-f <image>— attach an image. Repeatable. Images only — see video below.- No
-p? The prompt is read from stdin. Use stdin for prompts over a few KB (printf %s "$long" | gemini --dir .); argv has a ~1MB ceiling. - Long jobs are slow (a multi-image call can take minutes). Do not abort early.
Gemini has its own tools — exploit that
Inside OpenCode, Gemini runs as an agent with Read / Bash / Edit / network access, and
tool calls are auto-approved. So instead of stuffing files into the prompt, TELL IT TO
LOOK: gemini --dir /path/to/repo -p "Use your tools to read src/**/*.py and map the data flow. Cite file:line." This is the whole point of routing through OpenCode.
Verified: it ran curl -s https://example.com and Read hidden.txt unprompted.
For scraping, that means Gemini can drive curl, yt-dlp, jq, and page-parsing loops
itself. Give it the target and the output shape; let it iterate.
VIDEO — extract frames first, always
OpenCode cannot attach video: -f some.mp4 fails with Cannot read binary file
and Gemini replies with no visual input. The Gemini File API path (native video, FPS
sampling) needs a GEMINI_API_KEY and is NOT reachable through this OAuth route.
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 · -234 tokens per session aa7d6455e8e7
- 2mo ago First seen · 121 lines · 295 tokens per session scan A 25de3b4c0982
gemini is an agent published in the GitHub repository roarista/awesome-harness (1 stars, last pushed 4d ago), licensed MIT. It adds 61 tokens to every session and 1,669 once invoked, about $0.0002 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 2026-08-31.
gemini compared with agents from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| geminiroarista/awesome-harness | Agent | 1 | 4d ago | MIT | A | 61 | Claude Code |
| playwright-test-generatormicrosoft/playwright | Agent | 99k | today | Apache-2.0 | A | 151 | Claude Code |
| .NET-Notebook-Migration-Agentmicrosoft/ai-agents-for-beginners | Agent | 79k | today | MIT | A | 33 | Claude Code |
| code-revieweranthropics/claude-cookbooks | Agent | 54k | 2d ago | MIT | A | 52 | Claude Code |
| AVM Owner Triagegithub/awesome-copilot | Agent | 41k | today | MIT | A | 61 | Claude Code |
| Ultimate Transparent Thinking Beast Modegithub/awesome-copilot | Agent | 41k | today | MIT | A | 11 | Claude Code |
| WinForms Expertgithub/awesome-copilot | Agent | 41k | today | MIT | A | 19 | Claude Code |