The Agent Skills Directory

[SAFE]: The skill instructs the agent to analyze local code diffs and PRs to determine their impact. It encourages writing and running local test scripts to verify assumptions about code behavior, which is a standard development workflow and does not involve downloading or executing remote code from untrusted sources.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data such as pull requests and commit messages. While this presents a surface for indirect prompt injection (where instructions hidden in a PR might influence the agent's behavior), the skill focuses on verifying safety and explicitly reminds the user to 'strip anything private' before sharing results.
[DYNAMIC_EXECUTION]: The skill involves generating and executing local test scripts to prove specific code facts. This is part of the core functionality for code verification and is performed within the local project context rather than using data from the network.

blast-radius