callscript
a tool-calling script language for LLMs - the model authors ONE small script (plain JS, compiled to a validated, bounded, inert JSON plan - never executed) instead of calling tools one at a time. Adapter-based: mount AI SDK tools or plain { name, execute
No outside reports on file
Callscript is safer than live tool execution because plans aren't run automatically, but it's only as safe as the tools you connect and the prompts you give the model.
Based on public adoption and transparency signals as of Oct 28, 2026. Our opinion, not a code audit or security guarantee.
What it can access
- Local filesNo local folder access notedLOW
- Command executionNo shell command runner notedLOW
- Network reachLocal or standard API callsLOW
- CredentialsNo API token requiredLOW
- Package sourcenpm (callscript)PUBLIC
Inferred from this listing’s public package details and description. Check the project’s source before installing.
Install
npx -y callscriptAbout Callscript
Callscript is a scripting language that lets AI models write one small JavaScript-like plan instead of calling tools one at a time. The plan gets converted to a validated JSON structure and never actually runs—it just describes what should happen. You can plug in tools from the AI SDK or custom tools with a name and execute function.
Who it’s for
Developers building AI applications who want their models to plan multi-step tool use in a single pass rather than back-and-forth calls. Teams using Claude, other LLMs, or AI SDK tools who need safer, more efficient orchestration of external actions.
Alternatives
- Anthropic tool use: Claude's native tool_use feature for direct, structured function calls
- LangChain agents: Multi-step tool orchestration with built-in planning loops
- OpenAI Assistants: Managed AI agents with tool support and execution
- Replicant: Another MCP server for coordinating tool calls with planning
Setup
Check these steps against the project’s own instructions before running anything.
Install via npm:
npm install callscript
Then import and pass your tools to the script executor. Follow the project's README for adapter setup with your specific AI SDK or tool set.
Why this grade
Signal-based grade (1/7 points): no public source code found. This reflects adoption and transparency, not a code review.
Popularity and upkeep
- Weekly npm downloads
- 27.6k
- Listing last refreshed
- 2026-10-29
These are public popularity signals, not a maintenance check. Look at the repository’s commit history and open issues.
Security reports
No reports found for this package in our research (last checked 2026-10-29).
Risks to watch for
- Tool access: Callscript can reach any tool you mount to it—if a tool connects to APIs, databases, or file systems, the model can direct it there
- Prompt injection: A malicious prompt or data fed to the model could trick it into writing a script that calls tools in harmful ways
- Over-permissioned tools: If your mounted tools have broad permissions (delete files, access all APIs), the model might misuse them
- JSON plan leakage: The validated plan could be logged or exposed, revealing intent to call sensitive tools
- No sandboxing: The tool execution itself happens outside Callscript—your tool code runs with whatever permissions you gave it
How to lower the risk
Least-privilege permissions
- Mount only the tools the model actually needs
- Give each tool the minimum permissions required (read-only APIs, limited file paths)
Network isolation
- Run tools in a sandboxed environment or container when possible
- Use API keys with restricted scopes, not full credentials
Monitoring
- Log all generated scripts and which tools they call
- Alert if the model tries to call tools outside its normal pattern
Red flags to watch for
- Scripts that chain many tools in unexpected ways
- Repeated attempts to call tools that should be restricted
- Scripts accessing sensitive data or external systems without clear reason
Questions
How do I install Callscript?
Run npx -y callscript in your terminal, or add the server entry to your claude_desktop_config.json or .cursor/mcp.json as shown in the Setup section on this page.
Are there security reports about Callscript?
No reports found for this package in our research (last checked 2026-10-29). Callscript currently holds Grade D (High risk signals) based on public adoption and transparency signals. Key risk areas to watch include tool access, prompt injection, over-permissioned tools.
Is Callscript actively maintained?
Callscript has 27.6k weekly npm downloads, with its latest repository commit pushed on 2026-10-09. Our directory data was last refreshed on 2026-10-29. Check recent commit activity, release tags, and open issues before pinning it in production.
What are alternatives to Callscript?
Notable alternatives to Callscript include Anthropic tool use (Claude's native tool_use feature for direct, structured function calls), LangChain agents (Multi-step tool orchestration with built-in planning loops), OpenAI Assistants (Managed AI agents with tool support and execution), Replicant (Another MCP server for coordinating tool calls with planning). You can also compare Callscript side by side with other MCP servers in the Related tools section on this page.
README badge and team notes
Maintain or use Callscript? Show its live grade in your GitHub README, docs site, or run a check from the terminal before you install.
GitHub README Markdown:
[](https://agentarmorhq.com/skills/vercel-labs-callscript)HTML embed:
<a href="https://agentarmorhq.com/skills/vercel-labs-callscript"><img src="https://agentarmorhq.com/api/badge/vercel-labs-callscript" alt="AgentArmorHQ Risk Grade for Callscript" height="20" /></a>Terminal check:
curl -fsSL https://agentarmorhq.com/api/cli | node --input-type=module - check "vercel-labs-callscript"Share with your security lead
Need sign-off to use callscript? Copy this note for Slack, a ticket or a pull request.
### AgentArmorHQ Security Review Memo: callscript
- **Listing URL:** https://agentarmorhq.com/skills/vercel-labs-callscript
- **Type:** App connector
- **Risk Grade:** Grade D (High risk signals)
- **Outside Security Advisories on File:** 0
#### Permission & Blast-Radius Signals
- **Local files:** No local folder access noted
- **Command execution:** No shell command runner noted
- **Network reach:** Local or standard API calls
- **Credentials:** No API token required
- **Package source:** npm (callscript)
#### Recommended Hardened Setup
- **Command:** `npx -y --ignore-scripts callscript`
- **Controls:** Pin an exact version in `.mcp.json`, scope access to a single workspace folder, and use a least-privilege read-only token where supported.
_Based on public adoption and transparency signals as of Oct 28, 2026. Our opinion, not a code audit or security guarantee._Key facts
- Type
- App connector
- Install command
- npm install callscript
- Est. context load
- ~900 tokens / turn (~5 tool schemas, 0.45% of 200k context)
- Popularity
- 27.6k weekly npm downloads
- Last updated
- 2026-10-29
- Last graded
- 2026-10-28
Community
Discussions for Callscript
Ask questions or share configuration tips. Governed by our Community Rules & Terms.
No posts match this filter yet.
Be the first to ask a question or share a working tip.
Check before you install, in Claude
Your assistant answers Block, Review or No known issues for any package before it installs it.
Get this check inside Claude