The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill instructs the agent to create and run scripts or tests that execute the project's actual code to verify safety assumptions (Step 5). This is a core part of the skill's intended functionality for technical analysis.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data from version control systems, which represents a potential attack surface.
Ingestion points: The skill reads pull request content and commit history using gh pr view and git log (SKILL.md).
Boundary markers: No specific markers or delimiters are instructed for the external data.
Capability inventory: The skill uses gh and git tools and has the capability to write and execute scripts locally.
Sanitization: No explicit sanitization of the PR data or commit messages is described prior to analysis or script generation.
[COMMAND_EXECUTION]: The skill uses command-line tools such as gh (GitHub CLI) and git to perform its primary tasks of retrieving code and metadata for review.

blast-radius