Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed Ragie.ai RAG helper that sends user-selected documents and queries to Ragie for ingestion, retrieval, and document management.
Install this only if you intend for selected documents, URLs, queries, filters, and metadata to be processed by Ragie.ai. Keep RAGIE_API_KEY private, use partitions or filters for scoping, and review document IDs carefully before confirming deletion.
The skill exposes a document deletion command as a normal management action without requiring an explicit confirmation step or warning that the action is destructive and potentially irreversible. In an agent setting, this increases the chance of accidental or over-broad deletion from ambiguous user requests, causing data loss in the connected Ragie knowledge base.
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
payload["partition"] = partition
if metadata:
payload["metadata"] = metadata
resp = requests.post(f"{API_BASE}/documents/url", headers=headers, json=payload)
resp.raise_for_status()
return resp.json()
This code sends the user-provided query, and potentially metadata filter contents, to Ragie.ai via an HTTP POST request. While the module docstring identifies the service, there is no runtime warning, confirmation, or explicit disclosure that user-supplied content is being transmitted off-system.
The code accesses the RAGIE_API_KEY environment variable to authenticate outbound requests. Although this is functionally necessary and documented in the docstring, there is no explicit warning or disclosure to the user at execution time that credentials from the environment will be used for external network access.
No suspicious patterns detected.