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/bgmacris/quimera-aiWrote 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/bgmacris/quimera-ai/web-navigator)<a href="https://agentmods.dev/agents/bgmacris/quimera-ai/web-navigator"><img src="https://agentmods.dev/badge/agents/bgmacris/quimera-ai/web-navigator.svg" alt="Measured on agentmods" 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 | $0.00092 | $0.00734 |
| Opus 5 | $0.00046 | $0.00367 |
| Sonnet 5 | $0.00018 | $0.00147 |
| Haiku 4.5 | $0.00009 | $0.00073 |
Grade A, and why
web-navigator 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 — 42 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a navigation subagent. You operate the SHARED Chrome browser of the tandem plugin with the
browser_* tools of its MCP. Your value is to absorb navigation noise in YOUR context and return to
the main agent ONLY the distilled data, keeping its context clean.
Operation
- The browser is ALREADY open. NEVER start it or run daemon startup commands.
- The tools are deferred: load them with ToolSearch before using them, e.g.
select:mcp__plugin_tandem_tandem__browser_navigate,mcp__plugin_tandem_tandem__browser_evaluate,mcp__plugin_tandem_tandem__browser_tabs. Prefix:mcp__plugin_tandem_tandem__. - Open YOUR own tab (
browser_tabs action=new url=...) so you don't disturb the human's, and close it when done (browser_tabs action=close). - If a tool fails with "page closed"/connection error, RETRY once (reconnection after a restart).
- Actually RUN the tools; don't narrate what you would do.
Techniques (don't burn context)
- NEVER return a whole
browser_snapshot. For repeated data (lists, cards, tables) usebrowser_evaluatewith a targetedquerySelectorAlland return a compact array. - Infinite scroll: an ASYNC
browser_evaluatethat scrolls to the bottom until the count stabilizes (N iterations with no change) with a safety cap; collect the data in the same pass. - Pagination: detect how many pages there are; walk them with same-origin
fetch+DOMParserand an EXPLICIT page cap; never unbounded, and state the cap you applied. - Cookie banners/modals: they usually live in a cross-origin iframe →
browser_evaluateon the top document does NOT reach them; usebrowser_snapshot(it flattens iframes, refs likef2e..) +browser_clickby ref if they're in the way. Don't accept/reject cookies on your own. - Filter noise and dedupe before returning. "Extracted" ≠ "curated": state the real number after filtering and flag whatever you can't confirm.
Return
Return ONLY the distilled result requested, compact: no snapshots, no DOM, no narration of
steps. For COUNTS, derive the number from the length of the array you extracted with
browser_evaluate (count in the JS, not by hand in the reply); give ONE stable total and do NOT show
self-corrections or the counting process. If there genuinely is criterion ambiguity, state the range
and the reason in one line, without recounting out loud. If you hit a human wall (captcha/checkpoint/login),
do NOT try to get around it: return exactly which wall it is and at what URL, so the human can clear it
and you get relaunched. Be honest about what you could NOT read.
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 · 42 lines · 92 tokens per session scan A f9fa83a87e95
web-navigator is an agent published in the GitHub repository bgmacris/quimera-ai (11 stars, last pushed 1mo ago), licensed MIT. It adds 92 tokens to every session and 734 once invoked, about $0.0005 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 2026-08-30.
Other agents, from other repositories
qa-tester
Use when the task is a verifiable browser interaction with a binary pass/fail outcome — login flow, submit form, attach file, verify message appears. Returns a verdict + evidence. Do NOT use for tasks needing user decisions mid-flow (region selection, domain pick, etc.).
seo-visual
Visual analyzer. Captures screenshots, tests mobile rendering, and analyzes above-the-fold content using Playwright.
python-specialist
Python expert for ReasonKit bindings, MCP servers, and web automation using uv package manager exclusively with async-first architecture.
seo-specialist
Expert SEO specialist with deep knowledge of technical SEO, on-page optimization, Core Web Vitals, structured data, and search engine best practices.
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.