The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to configure a GitHub Action (marsiandeployer/vibers-action@v1) which downloads and executes code from a third-party repository during the CI/CD process.
[DATA_EXFILTRATION]: The skill includes a feedback mechanism that sends the repository URL and user-provided messages to an external endpoint (https://vibers.onout.org/feedback) via a curl command.
[COMMAND_EXECUTION]: The documentation provides a shell command (curl) to be executed by the agent or user to transmit data to the service's backend.
[INDIRECT_PROMPT_INJECTION]: The workflow establishes a feedback loop where an external service (Vibers) submits Pull Requests that the AI agent is then expected to review and merge. This creates a surface where external actors can provide instructions that the agent may follow during the PR review process.
Ingestion points: Pull Request content and commit messages generated by the marsiandeployer account.
Boundary markers: None specified; the agent is instructed to "review the PR, merge or comment" without explicit instructions to disregard embedded commands in the PR body.
Capability inventory: The agent has the capability to read files, comment on PRs, and potentially merge code depending on its broader toolset.
Sanitization: No sanitization or validation of the PR content is described before the agent processes it.

vibers-code-review