connect-declarative-schema
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of external data records (JSONL) and schema definitions (YAML) from S3 or GCS buckets, which serves as a potential attack surface for indirect instructions.
- Ingestion points: Data is retrieved from paths like
s3://<BUCKET_NAME>/data/*andgs://<BUCKET_NAME>/data/*. - Boundary markers: The skill does not explicitly instruct the agent to use delimiters or ignore embedded instructions when processing the ingested records.
- Capability inventory: The skill interacts with a Neo4j graph and allows querying via Cypher, providing the agent with visibility into the ingested data.
- Sanitization: Validation is focused on the presence of a mandatory
idfield and matching schema versions, rather than content sanitization. - [COMMAND_EXECUTION]: The skill provides templates for AWS IAM and GCP CLI commands to configure cloud resources.
- Evidence: SKILL.md includes commands such as
aws iam create-roleandgcloud storage buckets add-iam-policy-binding. These are standard administrative actions required for the skill's stated purpose and use placeholders for sensitive parameters.
Audit Metadata