github-open-source-lifecycle
Pass
Audited by Gen Agent Trust Hub on Aug 10, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill demonstrates high security maturity by incorporating explicit operating principles that prioritize secret management, least-privilege access, and evidence-based verification. It correctly identifies the boundary between local preparation and remote publication, ensuring that sensitive operations are gated by human authorization and automated checks.
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill references fetching data from external registries and GitHub surfaces (e.g., using
curlorwget) solely for the purpose of verifying that releases and deployments are correctly configured and publicly available. These operations target well-known, trusted services and do not involve the exfiltration of sensitive information. - [INDIRECT_PROMPT_INJECTION]: As a tool designed to audit and standardize external repositories, the skill naturally ingests untrusted data from manifests and CI configurations. This presents a theoretical attack surface for indirect prompt injection. However, the risk is mitigated by the skill's requirement for explicit authorization for mutating steps and its focus on established, ecosystem-native release controllers.
- Ingestion points: Repository manifests (
package.json,pyproject.toml, etc.), documentation, and GitHub API metadata. - Boundary markers: The skill relies on structured auditing workflows rather than specific prompt delimiters for external content.
- Capability inventory: Git operations (commit, tag, push), repository setting mutations, and Actions workflow dispatch.
- Sanitization: Instructions explicitly mandate the exclusion of secrets from logs, commands, and reports.
Audit Metadata