agentic-actions-auditor
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONREMOTE_CODE_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
gh apicommand to list and fetch the content of GitHub Actions workflow files from remote repositories. These operations are restricted to data retrieval from well-known GitHub services and are necessary for the skill's auditing functionality. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data from external GitHub workflows. It mitigates this attack surface by providing strict 'Bash Safety Rules' that forbid piping fetched content to interpreters or using it in shell command substitutions.
- [REMOTE_CODE_EXECUTION]: While the skill involves fetching remote content (composite actions and reusable workflows), it includes explicit methodology to prevent execution. It instructs the agent to treat fetched YAML content exclusively as data and provides specific warnings against using
eval,source, or piping tobash,python, ornode.
Audit Metadata