Outlit SDK

Integrate Outlit SDK for customer context for agents. Triggers when users need to add Outlit to any web framework (React, Next.js, Vue, Nuxt, Svelte, Angular...

This skill is an instruction-only integration helper that will scan your repository (package.json, Cargo.toml, lockfiles and common tracking/analytics files), suggest package-manager install commands, and ask you to add an Outlit key to your environment config. Before using it: (1) Do not paste private secret keys into chat — confirm whether the prompt asks for a public client key or a server secret. (2) Review any package-manager commands the agent suggests before running them. (3) Expect the agent to read many project files during detection; if your repo contains sensitive files, run the skill in a safe environment or limit its access. If you need clarification about which key to provide for server vs browser usage, verify with Outlit docs or your admin.

Type: OpenClaw Skill Name: outlit-sdk Version: 1.0.0 The skill's primary purpose of integrating an SDK is benign, utilizing standard package management and file modification operations. However, it is classified as 'suspicious' due to a significant prompt injection vulnerability vector identified in `SKILL.md`. The `Doc URL Map` includes an entry for `https://docs.outlit.ai/llms.txt`, which the agent is instructed to 'Fetch... as needed for implementation details.' The file name `llms.txt` strongly suggests it contains instructions intended for LLMs. This allows for dynamic loading of agent instructions from an external source, creating a supply chain risk where a compromised or malicious `llms.txt` could provide harmful directives to the agent, even if the `SKILL.md` itself does not contain explicit malicious instructions.