ci-cd-ops
Pass
Audited by Gen Agent Trust Hub on Mar 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands via
curl,jenkinsCLI, andghCLI tools to manage CI/CD pipelines. This is consistent with its stated purpose of DevOps automation. - [CREDENTIALS_UNSAFE]: The 'Error Handling' section suggests verifying tokens using
echo $GITHUB_TOKEN. This practice can inadvertently leak sensitive credentials into build logs or agent history. - [PROMPT_INJECTION]: The skill defines an attack surface for indirect prompt injection where user-provided repository URLs or job names are interpolated into shell commands.
- Ingestion points: Command line arguments such as
--repo,--job, and--workflowdefined inSKILL.md. - Boundary markers: No delimiters or boundary markers are specified to separate untrusted input from the command template.
- Capability inventory: Shell command execution via
curl,jenkins, andghtools, as well as file system access for configuration files. - Sanitization: No sanitization or validation logic is specified for these inputs, increasing the risk of command injection if the agent receives malicious inputs.
Audit Metadata