fix-sentry

SUSPICIOUS: The skill’s capabilities broadly match its stated purpose, and its GitHub CLI/tooling usage is consistent with official developer workflows. However, it grants an agent high-impact autonomous behavior: consuming untrusted Sentry content, editing code, running repository commands, pushing branches, and creating/updating PRs without user confirmation. The main risk is autonomous action and prompt-injection exposure, not obvious credential theft or malicious exfiltration.