The Agent Skills Directory

Indirect Prompt Injection (SAFE): The skill possesses a surface for indirect prompt injection as it processes external data (user feedback) and transmits it to a third-party service (GitHub).
Ingestion points: CLI arguments passed to scripts/create-issue.ts (e.g., --description, --actual).
Boundary markers: None observed in the interpolation logic.
Capability inventory: Uses gh for network operations and git for local repository metadata.
Sanitization: None; input strings are interpolated directly into the issue body.
Command Execution (SAFE): The skill executes gh and git commands using Bun.spawn with array arguments, which effectively prevents shell injection vulnerabilities. All tool usage is consistent with the skill's declared purpose in SKILL.md and uses only tools listed in the allowed-tools configuration.

stack-feedback