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 agentmods add skills/borda/vision-delivery/auth-setupnpx skills add Borda/vision-delivery --skill auth-setupgit clone --depth 1 https://github.com/Borda/vision-deliveryWrote 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/borda/vision-delivery/auth-setup)<a href="https://agentmods.dev/skills/borda/vision-delivery/auth-setup"><img src="https://agentmods.dev/badge/skills/borda/vision-delivery/auth-setup.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 | $0.00129 | $0.01007 |
| Opus 5 | $0.00064 | $0.00504 |
| Sonnet 5 | $0.00026 | $0.00201 |
| Haiku 4.5 | $0.00013 | $0.00101 |
Grade A, and why
auth-setup 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 4d 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 — 77 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Connect the bundled URL-only Roboflow MCP through the host-managed sign-in flow without asking for, reading, writing, or logging an API key. Keep useful local CV work moving when sign-in is declined or unavailable.
The repository .mcp.json contains only https://mcp.roboflow.com/mcp. Do not add headers, edit host configuration, create .env, or copy a remembered account/API-key recipe.
1 — Do useful work before auth
Authentication is needed only at a live Roboflow action seam. Continue local intake, media inspection, eval definition, architecture, artifact scaffolding, and economics work without blocking on account connection.
2 — Start host-managed sign-in
When a Roboflow operation is needed:
- Explain the specific read or action the connection enables and whether data may leave the machine.
- Invoke the least-invasive required Roboflow MCP read. The host should open the Roboflow sign-in/authorization flow.
- Ask the user to review the account and requested scope in the host UI. Never ask them to paste a token or key into chat.
- After authorization, retry one read-only operation and report only the connected workspace/account identifier needed for the task.
- Return to the active Sentinel workflow and its frozen acceptance gate.
If the user declines, continue with local evidence or a candid scaffold. Do not ask again in the same session unless they request a live platform action.
3 — Diagnose without guessing
Use read-only host inspection first:
- Claude Code:
claude plugin list --json - Codex:
codex plugin list
Confirm that sentinel@sentinel is enabled and the Roboflow MCP URL is present. Do not edit marketplace or MCP configuration automatically.
| Symptom | Action |
|---|---|
| Plugin missing or disabled | Reinstall with the two documented marketplace commands, start a new session, and retry. |
| Sign-in UI never appears | Preserve the host error; check host/network support against current official host and Roboflow guidance. Do not fall back to a guessed key header. |
| Unauthorized after sign-in | Ask the user to review the selected Roboflow account/scope, then retry one read-only operation. Do not request credentials in chat. |
Duplicate roboflow server after installing official skills |
Ask the user to retain one server registration for the session; do not mutate host config without explicit direction. Sentinel still delegates platform recipes upstream. |
| Host reports auth unsupported | State the exact limitation. Continue locally or use a separately verified upstream client flow; do not claim the plugin is authenticated. |
| MCP unavailable or schema changed | Stop at the platform seam, retain local work, and consult the installed official Roboflow skill/resource for the current recovery path. |
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.
- 4d ago First seen · 77 lines · 129 tokens per session scan A 1515bb1fd40c
auth-setup is a skill published in the GitHub repository Borda/vision-delivery (4 stars, last pushed 24d ago), licensed Apache-2.0. It adds 129 tokens to every session and 1,007 once invoked, about $0.0006 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
pipeline
Use when the user wants a feature idea taken end-to-end in one autonomous run — phrases like "run the whole pipeline", "take this feature from idea to finished branch", "brainstorm then build it autonomously", "do everything from idea to merged", "implement all phases without stopping". Triggers when they want…
craft
Use when a product idea is still vague and needs to become a clear definition of what to build — "let's craft an app like X", "help me define what I actually want", "clarify this idea before we plan it". Also use before planning or implementation when requirements, UX, domain behaviour, or technical preferences have…
boundedrelay-adaptive-sdd
Coordinate the optional BoundedRelay Spec Kit workflow with strict independent reviews, deterministic effort routing, bounded proposals, convergence, and handoff.
boundedrelay-setup
Verify and explain the local BoundedRelay connection used by Claude Code without changing provider models or enabling proposals implicitly.
prewalk
Arm a prewalk run: explore and plan in the root session, land the first verified edit, then persist a Stop checkpoint for handoff.
pw-go
Request the reviewed prewalk handoff and spawn exactly one routed Claude Task for the remaining todos.