sixel-email

1:1 email channel for agents — the agent can only email one address, and only that address can email the agent. Also handles the heartbeat (poll to prove you...

This skill appears to do only what it says: use the sixel.email API to send and receive messages for a single operator. Before installing, verify the following: 1) Confirm the SIXEL_API_TOKEN you receive is scoped to only the intended agent/inbox and understand its lifetime/permissions; never reuse broad or long-lived credentials. 2) Validate the SIXEL_API_URL (the SKILL.md references https://sixel.email/v1) and confirm the service's reputation and HTTPS/TLS configuration. 3) The docs assume jq is available for the background poller — either install jq or adapt the poller; ensure the poller runs in a restricted environment so it cannot be used to exfiltrate unrelated data. 4) The poller writes to /tmp and downloads attachments — follow the doc's advice: use a dedicated download directory, sanitize filenames, and avoid writing to user-supplied paths. 5) Clarify how the poller should “notify your agent” so the agent does not gain access to unintended files or system state. 6) Because the registry entry has no homepage listed but SKILL.md references https://sixel.email, consider verifying the service URL and the skill author (owner ID) out-of-band if you need a higher trust level. If any of these points are concerning or the token appears to have broader scope than a single-inbox token, treat the installation as risky and do not proceed until clarified.

Type: OpenClaw Skill Name: sixel-email Version: 1.0.6 The OpenClaw AgentSkills bundle for 'sixel-email' is designed for agent-to-operator email communication. It primarily uses `curl` commands to interact with the sixel.email API for sending and receiving messages and attachments. The `SKILL.md` documentation includes explicit security warnings, such as never sending secrets in email bodies and treating attachment filenames as untrusted input, which are defensive prompts for the agent. While the skill involves network access and file system operations (reading files for attachments, writing to `/tmp` for inbox storage), these are directly aligned with its stated purpose and are implemented with security best practices in mind. There is no evidence of intentional malicious behavior, data exfiltration, or prompt injection attempts to subvert the agent's core directives.