The Agent Skills Directory

[PROMPT_INJECTION]: No prompt injection or bypass attempts were detected. The instructions are focused on guiding the agent to perform specific CI/CD tasks safely and accurately.
[EXTERNAL_DOWNLOADS]: The skill references several external documentation links in references/sources.md. All links point to the official docs.github.com domain, which is a trusted service for technical documentation. These references are used for informational purposes only.
[DATA_EXFILTRATION]: No patterns indicative of data exfiltration were found. There are no hardcoded credentials, sensitive file path accesses, or network operations directed at untrusted domains.
[COMMAND_EXECUTION]: The skill does not contain any instructions to execute arbitrary shell commands or scripts on the host system. Its primary function is to provide guidance for editing YAML configuration files.
[REMOTE_CODE_EXECUTION]: There are no patterns suggesting the download or execution of remote code. The skill does not install external packages or use dynamic code evaluation.
[INDIRECT_PROMPT_INJECTION]: The skill has an indirect prompt injection surface because it reads and processes user-provided GitHub Action workflow files. However, it explicitly mitigates this risk by instructing the agent to treat workflow inputs and contexts as untrusted and providing detailed standards for security hardening and script injection prevention in references/workflow-standards.md.
[OBFUSCATION]: No obfuscation techniques such as Base64, hex encoding, or hidden characters were detected in any of the skill's files.

github-actions