Musallat Bot

Otonom, pasif-agresif yazılımcı bot, teknik hatalara tahammülsüz, kibarlığı reddeden ve gereksiz açıklamalara sert yanıt veren kıdemli programcı.

This skill appears to be a small Gemini-based persona bot, but there are inconsistencies you should resolve before installing. Key points to consider: - Do not assume the API key shown in SKILL.md is safe to use: the bracketed string (starting with 'AIzaSy') looks like a Google API key and may be a leaked secret or example. Treat it as compromised until verified. If it is yours, rotate it immediately. - The Python code expects GEMINI_API_KEY in the environment but the skill manifest does not declare this. Ask the author to explicitly document required env vars and not embed real keys in docs. - The skill imports google.generativeai but provides no install steps; ensure your environment has that package and review the package provenance before installing third-party libraries. - Because the skill will send prompts to an external model using your key, using it can incur cost and will transmit user-provided prompts to Google. Prefer creating a dedicated, limited-scope API key for this skill and monitor usage. Recommended actions before enabling: 1) Ask the publisher to remove the embedded API key from SKILL.md (or confirm it's a harmless placeholder). If it was a real key, rotate it. 2) Require the author to add GEMINI_API_KEY to the declared required env vars and document needed permissions/quotas. 3) Run the skill in an isolated environment and audit outbound traffic to verify endpoints are only Google generative API endpoints. 4) If you cannot verify the key or author, treat the skill as untrusted and do not provide any sensitive credentials.

Type: OpenClaw Skill Name: musallat-bot Version: 1.0.1 The `skill.md` file contains direct instructions to the OpenClaw agent to adopt a specific persona, which constitutes a form of prompt injection, even though its objective is not overtly malicious. Additionally, `skill.md` hardcodes an API key (`AIzaSyBxfb-8s5TsOVvr55_E5lDbilpVLoSwIj8`), which is a credential exposure, even if intended for the skill's own use. While the Python code attempts to retrieve the API key from an environment variable, the hardcoded key in the markdown file represents a security flaw and a risky capability.