The Agent Skills Directory

[SAFE]: The skill performs standard local file operations, reading repository source code (routers, services, schemas) and documentation (ADRs) to generate or update a domain language file (CONTEXT.md).
[SAFE]: No network access or data exfiltration mechanisms were found. The skill does not use tools like curl or wget, nor does it reference external URLs.
[SAFE]: There is no evidence of obfuscation, hardcoded credentials, or attempts to access sensitive system files (e.g., SSH keys or environment secrets).
[SAFE]: The skill's logic is purely instructional and does not include prompt injection patterns intended to bypass safety guardrails or escalate privileges.

evanflow-glossary