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 agents/callstack/agent-device/web-backendgit clone --depth 1 https://github.com/callstack/agent-deviceWhat 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 | $0.00000 | $0.00459 |
| Opus 5 | $0.00000 | $0.00230 |
| Sonnet 5 | $0.00000 | $0.00092 |
| Haiku 4.5 | $0.00000 | $0.00046 |
Grade A, and why
web-backend 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 yesterday.
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.
Runs shell commandslowCapability
Expected in a hook, worth knowing in a rule or an instructions file.
`child_process.spawn` refuses without a shell (CVE-2024-27980 hardening), and a shell would What it actually says
Web Backend
Web automation uses a managed agent-browser backend as an implementation detail.
- Runtime web commands resolve the backend only from the state-dir managed install at
tools/agent-browser/<version>. - Normal
--platform webcommands do not install the managed backend on first use. If the backend is missing, they fail with a setup hint. - Use
agent-device web setupbefore first web automation and in CI/sandbox bootstrap steps. - Use
agent-device web doctorto run the backend health check. - The managed install respects
--state-dir/AGENT_DEVICE_STATE_DIR. - Web automation requires Node 24+ while the rest of agent-device keeps its Node 22 baseline.
- Every backend call spawns the package's declared
binentry with the current Node runtime, never thenode_modules/.binconsole shim: Windows ships that shim as.cmd, whichchild_process.spawnrefuses without a shell (CVE-2024-27980 hardening), and a shell would reintroduce argument-quoting hazards.runManagedAgentBrowseris the only path that executes the backend. Setup spawnsnpmfrom PATH as before, except on Windows, where a barenpmis not spawnable and npm's ownnpm-cli.jsruns under the current Node instead.
Default first-run flow:
agent-device web setup
agent-device open "https://example.com" --platform web
agent-device snapshot -i --platform web
agent-device viewport 1280 900 --platform web
agent-device screenshot ./artifacts/web-full.png --platform web --fullscreen
agent-device network dump 25 --platform web
agent-device close --platform web
Do not document direct agent-browser commands as agent-device features. Web network dump is the
narrow exception: it adapts agent-browser network requests to the existing agent-device network
evidence shape. Browser-specific network routing/interception/HAR, CDP, React web, tabs, downloads,
auth vaults, and profiling stay out of the minimal web surface until there is an explicit
agent-device command design for them.
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.
- yesterday First seen · 35 lines · 0 tokens per session scan A bf1ff67eddea
web-backend is an agent published in the GitHub repository callstack/agent-device (4,300 stars, last pushed today), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 459 tokens. A static security scan graded it A with 1 finding (runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
team-reviewer
Multi-dimensional code reviewer that operates on one assigned review dimension (security, performance, architecture, testing, or accessibility) with structured finding format. Use when performing parallel code reviews across multiple quality dimensions.
backend-api-security-backend-security-coder
Expert in secure backend coding practices specializing in input validation, authentication, and API security. Use PROACTIVELY for backend security implementations or security code reviews.
agent-orchestration-context-manager
Elite AI context engineering specialist mastering dynamic context management, vector databases, knowledge graphs, and intelligent memory systems. Orchestrates context across multi-agent workflows, enterprise AI systems, and long-running projects with 2024/2025 best practices. Use PROACTIVELY for complex AI…
basic-agents
A basic agent uses a predefined strategy with a simple execution flow that works for most common use cases. It accepts a string input (a question, request, or task description) and sends this input to the configured LLM. The LLM may decide to call provided tools. The agent will execute the tools and send the results…
explainer
Code explanation and architecture walkthroughs. Helps developers understand complex code and systems.
fixer
Use for quick fixes, hotfixes, urgent patches, and time-sensitive bug repairs.