The Agent Skills Directory

[Secure Resource Handling]: The instructions explicitly forbid hardcoding or logging sensitive credentials, directing the agent to use environment variables for secret management. This is a recommended security practice for handling authentication data.
[Path Validation and Access Control]: The skill requires validation of connector IDs and file paths to ensure that read and write operations are restricted to the expected directory (~/.openwiki/connectors/). This helps prevent path traversal vulnerabilities.
[Input Ingestion Surface]: While the skill involves creating connectors that ingest external data—a potential surface for indirect prompt injection—it provides clear guidance on treating external servers as read-only and validating manifests. These measures align with a robust security posture for data processing.

write-connector