Command-line tool to manage contacts, tasks, notes, deals, events, and communications via the Follow Up Boss API.

This skill appears to be a straightforward Follow Up Boss CLI, but there are two actionable concerns you should resolve before installing or using it: - Metadata mismatch: The registry entry declares no required environment variables or primary credential, yet SKILL.md and the code require an API key (FUB_API_KEY). Treat the API key as sensitive: do not paste it into public places and prefer injecting it via secure environment mechanisms. - Untrusted source/homepage: The skill lists no homepage or known source. If you plan to use it, review the fub.js file yourself (it is included) and confirm it does only what you expect (calls api.followupboss.com). Consider running it in a restricted environment or container first. Additional notes: - The code also supports FUB_SYSTEM_KEY and FUB_SYSTEM_NAME (not documented in SKILL.md); if you use system integrations, ensure you understand those headers. - Ensure your Node runtime provides global fetch or run with a Node version that supports it, otherwise inspect/modify the script to include a fetch polyfill you trust. - If you proceed, provide the minimum-scoped API key possible, and rotate the key if it is accidentally exposed. If the publisher can update the registry metadata to list FUB_API_KEY as a required credential and provide a homepage/source repository, that would remove the main inconsistency and increase confidence.

Type: OpenClaw Skill Name: follow-up-boss Version: 1.0.0 The skill bundle is classified as suspicious due to several input validation vulnerabilities in `fub.js`. Specifically, user-supplied arguments are directly concatenated into URL query strings (e.g., `people?${args[1] || ''}`) and URL paths (e.g., `people/${args[1]}`) without proper sanitization or encoding, except for the `search` command which uses `encodeURIComponent`. This creates a URL injection risk and potential for ID manipulation if a malicious user provides crafted input. There is no evidence of intentional malicious behavior such as data exfiltration, backdoors, or prompt injection against the agent in `SKILL.md`.