The Agent Skills Directory

[Data Exposure & Exfiltration] (SAFE): The skill does not contain any hardcoded credentials, sensitive file path access, or network exfiltration logic. It operates solely on the context of provided GitHub issues.
[Remote Code Execution] (SAFE): No remote script execution, package installation, or dynamic code generation was found. The skill is entirely descriptive and instruction-based.
[Indirect Prompt Injection] (SAFE): The skill is designed to ingest external data (GitHub issue descriptions and comments), which is a known attack surface. However, the risk is mitigated as the primary purpose of the skill is analysis rather than execution.
Ingestion points: Processes external GitHub issue descriptions, labels, and comments via the Read Issue step in SKILL.md.
Boundary markers: No specific boundary markers are defined to isolate untrusted data.
Capability inventory: The skill's capabilities are limited to scoring and reporting; no dangerous tools (filesystem write, shell execution) are utilized.
Sanitization: No sanitization logic is present, but the risk level is considered SAFE given the intended use case.
[Obfuscation] (SAFE): No obfuscated content, Base64 encoding, or hidden characters were detected in the skill files.

issue-triage