zworkforce-skill-registry-governance

name	zworkforce-skill-registry-governance
description	Govern the zWorkforce signed remote skill registry, including manifest validation, HMAC signature verification, HTTPS host allowlisting, redirect handling, and install audit trail.

zWorkforce Skill Registry Governance

Only let trusted, verifiable skills into a tenant's agent runtime.

Workflow

1. Verify the skill manifest validates against the schema (DNS-like id, bounded version string, well-formed allowed_tools, bounded system_prompt_append) before considering installation.
2. Verify the manifest's HMAC signature against the configured signing key when signature enforcement is required; reject unsigned or invalid signatures rather than installing with a warning.
3. Confirm the registry URL is HTTPS, has no embedded userinfo, and resolves to an allowlisted host and a public (non-private, non-loopback, non-link-local) address.
4. Re-validate host allowlisting and public-address checks on every redirect hop, and cap redirect depth.
5. Confirm every install is recorded in the audit trail with source URL, version, and actor before treating a skill as active for a tenant.

References

• zworkforce/skill_registry.py
• zworkforce/skills.py
• docs/SECRET-MANAGEMENT.md
• docs/THREAT-MODEL.md

Output

Report manifest validation result, signature verification status, host/redirect checks performed, and audit trail confirmation for each install reviewed.