The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill uses the exclamation-backtick syntax to execute git rev-parse and gh auth status when the skill is loaded into the agent's context. These commands are benign and are used solely to populate the 'Current State' section with information about the current git branch and the user's GitHub authentication status.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from GitHub pull request comments (e.g., bot reviews) and possesses the capability to perform state-changing operations like merging branches. This creates a surface for indirect prompt injection if malicious instructions were embedded in a PR comment. However, the skill explicitly instructs the agent to verify branch protection rules, required reviews, and CI status before merging, which mitigates the risk of unauthorized or malicious actions. * Ingestion points: GitHub PR metadata, status check results, and pull request comments (SKILL.md). * Boundary markers: The skill does not define specific prompt delimiters for external content but groups findings by bot type. * Capability inventory: gh pr merge, gh pr delete-branch, gh pr view (SKILL.md). * Sanitization: The skill relies on the underlying GitHub CLI's logic and the repository's server-side branch protection settings.

gh-dash