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 Darwin-Agent/Car-bench-TRACE --skill poi-searchgit clone --depth 1 https://github.com/Darwin-Agent/Car-bench-TRACEWrote 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/darwin-agent/car-bench-trace/poi-search)<a href="https://agentmods.dev/skills/darwin-agent/car-bench-trace/poi-search"><img src="https://agentmods.dev/badge/skills/darwin-agent/car-bench-trace/poi-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/darwin-agent/car-bench-trace/poi-search"><img src="https://agentmods.dev/badge/skills/darwin-agent/car-bench-trace/poi-search.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 | $0.00210 | $0.04470 |
| Opus 5 | $0.00105 | $0.02235 |
| Sonnet 5 | $0.00042 | $0.00894 |
| Haiku 4.5 | $0.00021 | $0.00447 |
Grade A, and why
poi-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 10d 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 — 96 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Find a POI of a category at a place
The user wants to locate places of a given category — a restaurant, a hotel, a charger, … — at some location, or along their route. You resolve the location, search POIs of that category, and present the candidates for the user to choose from. Sometimes the category is left open; sometimes the category-search parameter is missing. The same method handles all of these. This skill finds POIs only; dialing one is the place-phone-call skill.
When this applies
"Find me a restaurant in ", "is there a at my destination?", "find somewhere to eat at my next stop", "find somewhere to eat where I'm headed", "look up nearby", "find a along my route around km." The place may be a named city, or described relative to the active route ("at my destination", "at my next stop"); the category may be named or left to the user's known preference.
Tools
get_current_navigation_state({detailed_information:true})— read the active route when the place is relative to the route. For "my destination" take the final waypoint's location id; for "my next stop" take the first upcoming intermediate waypoint (the one right after the current start), which on a multi-stop route is not the final destination. Don't ask where the destination/stop is — read it.get_location_id_by_location_name({location})— resolve a named place to a location id. Don't read navigation state for a named place. Search the place name as the user gave it; strip only an appended country/qualifier suffix (", Germany", "City Center"), never the distinctive part of the name itself — "Andorra la Vella" is the city's own name, so don't reduce it to "Andorra". If a lookup returns not-found, retry with the user's full name before declaring the place unavailable; only then ask for a nearby searchable place. Always use the id this tool returns — never guess or fabricate a location/POI id, and don't fire a dependent route/search call with a guessed id in the same parallel batch as the resolution; wait for the real id first.get_user_preferences({...})— read the user's preferred category of place; this is what resolves an unstated category.search_poi_at_location({location_id, category_poi, filters})— list POIs of a category at a location.filtersis optional:any::currently_openkeeps only places open right now; charging filters (charging_stations::has_available_plug,charging_stations::has_dc_plug) narrow to stations with a free plug or a DC plug. Apply a filter only when the user's request actually demands it. Itscategory_poiargument may be the removed parameter; without it the call errors.search_poi_along_the_route({route_id, at_kilometer, category_poi, filters})— POIs along a route at a kilometre mark (10 km radius).at_kilometeris required for charging stations; pick it from the user's stated mark, convert a stated time into a route distance withconvert_route_distance_and_timefirst, or — when no mark is given but a charging preference exists — derive it from that preference. Samefiltersas above. May also be unavailable in a given setup.get_distance_by_soc({initial_state_of_charge, final_state_of_charge})— how far the car can drive between two charge levels. Use it to turn an SOC-based charging preference (e.g. "charge at 15% left") into the route distance at which to search for a charger.
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.
- 10d ago First seen · 96 lines · 210 tokens per session scan A c2a15d6bfcf8
poi-search is a skill published in the GitHub repository Darwin-Agent/Car-bench-TRACE (8 stars, last pushed 7d ago), licensed MIT. It adds 210 tokens to every session and 4,470 once invoked, about $0.0011 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 skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
insight-error-page
Write or audit an insight-kind error page for the Next.js dev overlay. Use when creating a new errors/ .mdx page, auditing an existing one, or checking that a page matches the framework fix cards. Covers page structure, title alignment, FixCard cards with Copy prompt button, code snippets, terminology verification…
next-cache-components-optimizer
Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…
next-partial-prefetching-adoption
Turn on Partial Prefetching in a Next.js app and work through the insights it surfaces. Use when the user wants to enable or adopt Partial Prefetching, flip the partialPrefetching flag, opt routes in with export const prefetch = 'partial', audit Link prefetch={true} behavior, preserve existing prefetched UI with…