The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documentation includes a JavaScript example for a webhook handler (webhook-handler.js) that uses execSync to run shell commands. It interpolates the event.comment.body variable directly into the command string without sanitization. This pattern allows for potential command injection if a user provides a comment containing shell metacharacters such as semicolons, backticks, or pipes to execute arbitrary commands on the system hosting the webhook.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from GitHub pull requests, creating a vulnerability surface for indirect prompt injection.
Ingestion points: Data enters the agent context via gh pr view (fetching diffs and descriptions) and through PR comment webhooks as shown in the integration examples.
Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions for the data being processed by the underlying review agents.
Capability inventory: The skill utilizes gh CLI for repository management and npx for executing review agents, which provides a capability tier including repository modification and network access.
Sanitization: The provided code examples lack input validation or escaping for external content interpolated into commands.
[EXTERNAL_DOWNLOADS]: The skill makes use of npx to execute the ruv-swarm package. This resource is associated with the author's toolset and is used to coordinate the parallel review agents.

github-code-review