publish-project-to-github
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from local project directories to generate repository content, README documentation, and audit reports.
- Ingestion points: The agent reads the local project directory, file contents, and directory structures as described in
SKILL.md(Steps 1, 2, and 4) and the audit scriptscripts/audit_public_project.sh. - Boundary markers: While the skill lacks explicit multi-modal delimiters for all interpolated data, it includes strong instructional guidance for the agent to audit and verify data before publication.
- Capability inventory: The skill possesses the ability to execute shell commands (
git,gh,bash,python3), write to the local filesystem (README generation), and perform network operations via the GitHub CLI. - Sanitization: The skill proactively employs
scripts/audit_public_project.shwhich usesripgrep(rg) to scan for secrets, private keys, and absolute system paths to block the publication of sensitive information.
Audit Metadata