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 TasiTech/tasi-harness --skill deep-searchgit clone --depth 1 https://github.com/TasiTech/tasi-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/skills/tasitech/tasi-harness/deep-search)<a href="https://agentmods.dev/skills/tasitech/tasi-harness/deep-search"><img src="https://agentmods.dev/badge/skills/tasitech/tasi-harness/deep-search/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/tasitech/tasi-harness/deep-search"><img src="https://agentmods.dev/badge/skills/tasitech/tasi-harness/deep-search.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00105 | $0.02611 |
| Opus 5.5 | $0.00042 | $0.01044 |
| Sonnet 5.5 | $0.00021 | $0.00522 |
| Haiku 4.5 | $0.00011 | $0.00261 |
Grade A, and why
deep-search 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 2mo 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 — 164 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Deep Search
Use this skill to run browser-backed research across search engines, collect candidate result links, de-duplicate sources, open the most relevant pages, and gather browser_snapshot / browser_extract evidence for downstream answering.
This workflow works in visible or headless external CDP mode. In headless mode, rely on browser_snapshot, browser_extract, browser_screenshot, and page state instead of visual browser inspection.
Search Engines
Choose engines by task and network conditions:
- Broad/default: Bing
https://www.bing.com/search?q=<query>&count=10 - International: Google
https://www.google.com/search?q=<query>&num=10, Google HKhttps://www.google.com.hk/search?q=<query>&num=10, DuckDuckGohttps://duckduckgo.com/html/?q=<query>, Bravehttps://search.brave.com/search?q=<query> - Chinese/China-local: Baidu
https://www.baidu.com/s?wd=<query>, Bing CNhttps://cn.bing.com/search?q=<query>&ensearch=0, 360https://www.so.com/s?q=<query>, Sogouhttps://www.sogou.com/web?query=<query>, WeChat via Sogouhttps://weixin.sogou.com/weixin?type=2&query=<query>, Toutiaohttps://so.toutiao.com/search?keyword=<query>, Quarkhttps://quark.sm.cn/s?q=<query>&from=smor&safe=1 - Finance/news direct leads when search results are thin: Sina Finance, Eastmoney, Tencent Finance, official exchanges/regulators, company IR pages, government/academic sources.
Use more than one engine when coverage matters, results look thin, the topic may differ across regions/languages, or one engine is blocked. For Chinese-language or China-local queries, try Bing CN plus Baidu/Sogou/360/Toutiao. For international queries, try Bing plus Google or DuckDuckGo/Brave. Treat search-engine AI summaries as leads, not evidence.
Robust Browsing Pattern
- Open each search URL with
browser_open, thenbrowser_waitfor load state, result text, or known result selectors. - Use a normal desktop viewport with
browser_viewportwhen responsive/mobile layouts hide links. - After initial load, use one small
browser_scrollto trigger lazy results, then wait briefly before extracting. - If an engine returns zero usable organic links, captcha/interstitial content, consent-only content, or regional blocking, record it and switch engines. Do not try to bypass captchas, paywalls, login walls, or explicit access controls.
- Prefer visible or headless CDP browser tools over raw network fetching because search pages often need JavaScript, redirects, and accessibility-aware link inspection.
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.
- 2mo ago First seen · 164 lines · 105 tokens per session scan A 548ab1c0939d
deep-search is a skill published in the GitHub repository TasiTech/tasi-harness (11 stars, last pushed 1mo ago), licensed MIT. It adds 105 tokens to every session and 2,611 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 2026-08-30.
Other skills, from other repositories
interactive-login
Log a CLI in with a browser/device-code flow (aws sso, gh, glab, gcloud, …) and save the result to the keychain so later commands can request it through execute.credentials.
pinchtab-mcp
Use this skill when a task requires browser automation through PinchTab's MCP server connected to a remote browser instance. Covers navigation, element interaction, data extraction, form filling, multi-step flows, and session management via MCP tools.
azure-messaging-webpubsub-java
Build real-time web applications with Azure Web PubSub SDK for Java. Use when implementing WebSocket-based messaging, live updates, chat applications, or server-to-client push notifications.
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.
open-design-browser
Inspect and exercise local HTML design artifacts and library prototype previews with the existing agent-browser CLI, including screenshots, interaction checks, and rendered DOM snapshots. Use when rendered evidence helps author or review a design; excludes browser installation, remote services, and unrelated browser…
browser_cdp
A health check for Python code that produces a score, letter grade, and measures of issues such as complexity, duplication, dead code, dependencies, and architecture.