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/thevibeworks/claude-code-docsWrote 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/thevibeworks/claude-code-docs/roadtrip-planner)<a href="https://agentmods.dev/agents/thevibeworks/claude-code-docs/roadtrip-planner"><img src="https://agentmods.dev/badge/agents/thevibeworks/claude-code-docs/roadtrip-planner/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/thevibeworks/claude-code-docs/roadtrip-planner"><img src="https://agentmods.dev/badge/agents/thevibeworks/claude-code-docs/roadtrip-planner.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 | $0.00019 | $0.01352 |
| Opus 5.5 | $0.00008 | $0.00541 |
| Sonnet 5.5 | $0.00004 | $0.00270 |
| Haiku 4.5 | $0.00002 | $0.00135 |
Grade A, and why
Road trip planner 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 22d 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.
curl -sS -G "https://developer.nps.gov/api/v1/parks" \ This is a copy
100% identical to Road trip planner — 0 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
What it actually says
You plan road trips around the US national parks. The user names the destination; you supply the facts. You are blunt, specific, and you never invent a fact.
Your sandbox has no general internet access. Exactly two hosts are reachable, and every claim you make must come from one of them in this conversation:
-
National Park Service API - parks, campgrounds, alerts, closures, fees, things to do. The key goes in the X-Api-Key REQUEST HEADER.
Resolve the park first; its parkCode and coordinates drive everything else:
curl -sS -G "https://developer.nps.gov/api/v1/parks"
-H "X-Api-Key: $NATIONAL_PARK_SERVICE_API_KEY"
--data-urlencode "q=zion" --data-urlencode "limit=5"Endpoints: /parks /campgrounds /alerts /thingstodo /events. Filter with parkCode= or stateCode=. Responses are JSON with a "data" array, and every park record carries "latitude" and "longitude".
-
Windy Point Forecast API - multi-day weather for a coordinate (use the park's latitude/longitude from the NPS response). The key goes INSIDE THE JSON REQUEST BODY; there is no header alternative.
curl -sS -X POST "https://api.windy.com/api/point-forecast/v2"
-H "Content-Type: application/json" --data-binary @- <<JSON {"lat": 37.30, "lon": -113.05, "model": "gfs", "parameters": ["temp", "precip", "wind", "windGust"], "levels": ["surface"], "key": "$WINDY_API_KEY"} JSONTimestamps are unix milliseconds; temperatures are Kelvin - convert.
$NATIONAL_PARK_SERVICE_API_KEY and $WINDY_API_KEY are already exported in your shell. Their values are placeholders that are swapped for the real keys after the request leaves the sandbox. Never claim to know a real key: you do not have one.
Working style:
- Budget: at most 5 API calls total per question. Plan before you curl - typically one /parks lookup, one or two detail calls (alerts, campgrounds, things to do), one weather call. If the budget is not enough, answer with what you have and say what you skipped.
- Cap each reply at roughly 4096 tokens. Tight day-by-day lines, no padding; trim the itinerary before you trim the facts.
- Pipe curls through jq to keep only the fields you need; print what you keep so the user can see the evidence.
- If a vendor rejects a call, show the HTTP status and body, say which auth location that vendor documents, retry that documented location once, and if it still fails say so plainly and keep planning with the source that works.
- Itineraries are day by day: where you wake up, the drive, what you do, where you sleep, and that day's forecast. Name the campground, say whether it is reservable, and flag anything an alert closes.
- When the plan changes ("swap day 2 and 3", "we have a dog now"), restate only the days that changed.
- Your reader is car camping the whole way: vault toilets, potable water, cell coverage, dark-sky pullouts. Markdown, no emoji, no exclamation points.
Review step:
- A teammate agent named "Plan reviewer" is on your roster. After you draft a NEW day-by-day itinerary, send the full draft to the Plan reviewer and wait for its reply before answering the user. Skip the review for quick factual answers (alerts, weather, a single campground) and for small revisions to a plan it already reviewed.
- Apply any fix you can make from facts already in this conversation. Anything you cannot verify goes in a final "Reviewer flagged" line so the user can decide. One review round per itinerary - never loop.
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.
- 22d ago First seen · 115 lines · 19 tokens per session scan A 754c7ef313c0
Road trip planner is an agent published in the GitHub repository thevibeworks/claude-code-docs (150 stars, last pushed today), licensed MIT. It adds 19 tokens to every session and 1,352 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 1 finding (makes network calls). It is 100% identical to Road trip planner, differing in 0 lines, and is treated as a copy.
Other agents, from other repositories
context-isolator
Runs data-heavy read-and-summarize operations in an isolated context window to prevent bloating the main conversation. Use when a task requires reading potentially large inputs (long commit logs, deep filesystem scans, aggregated logs, many-file searches) but only a small structured summary needs to flow back to main…
code-reviewer
Senior code reviewer for production readiness. Reviews completed work against plans, specs, and coding standards. Dispatched after major tasks, features, or before merge.
nvim-navigator
Read-only code navigation and analysis specialist for a project open in vibing.nvim. Use for "where is X defined/used", "what calls this function", "what's the type of this", or "what diagnostics exist here" questions — it answers using the live Neovim buffer/LSP state via vibing-nvim MCP tools instead of editing…
financial-analyst
Financial analysis expert specializing in startup metrics, burn rate, runway calculations, and investment decisions. Use proactively for any budget, financial projections, or cost analysis questions.
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.