OpenClaw gateway and toolchain test harness for PlatPhormNews.
Claws discovers MCP tools, APIs, schemas, workflows, and services, then lets humans, developers, IDEs, CLI tools, MCP clients, and future runtimes inspect, plan, validate, dry-run, trace, and hand off toolchains safely.
Protected execution, memory writes, workspace persistence, report publishing, Sandbox, BrowserOps, and Evals triggers require PLATPHORM_API_KEY.
Discovers tools, schemas, services, policies, and workflows, then turns them into testable agent plans.
Build from Start, MCP Tool, API Call, Spec Validate, Sandbox, BrowserOps, Evals, Trace, and Report nodes.
Public reads and dry-runs stay open. Mutations, protected tool calls, provider credentials, and private evidence require the shared key.
Every plan includes run id, trace id, redacted inputs, artifact links, degraded states, and a reproduction command.
Toolchain Testing Lifecycle
The main loop is validation first, execution second, evidence always. Claws can plan every step publicly, but protected actions remain queued until the operator supplies PLATPHORM_API_KEY.
Integration Matrix
Configured integration records are public. Add live probes from the API when operator evidence is needed.
What can Claws safely do?
Claws is published by PlatPhorm News Network as an operator cockpit for inspectable agent plans. It can discover registered tools, validate schemas, assemble toolchains, generate deterministic dry-runs, and link evidence to traces. It does not claim a provider or tool executed when no handler ran. Public reads expose registry and planning state; workflow execution, persistence, reporting, test triggers, remediation, and administrative changes require the shared platform credential.
- Registry-backed tools and schemas
- Human approval for protected actions
- Honest unavailable and degraded states
Protocol behavior follows the Model Context Protocol specification, and request correlation follows the W3C Trace Context recommendation.
Published by PlatPhorm News Network ·