The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): The skill uses git and gh CLI for repository management and status checks. All shell commands are standard for development and do not exhibit high-risk patterns.
[DATA_EXFILTRATION] (SAFE): External communication is limited to the user's configured GitHub environment. No unauthorized data access or credential harvesting was found.
[PROMPT_INJECTION] (SAFE): The skill possesses a surface for indirect prompt injection as it processes GitHub issue and PR content (Ingestion points: gh pr list, gh issue view in SKILL.md; Boundary markers: Absent; Capability inventory: Bash, Write, Edit tools; Sanitization: Absent). However, this is an inherent part of the tool's intended use case and the workflow includes specific isolation checks that mitigate risks, resulting in a SAFE assessment.

pr-implement