qdrant
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [Indirect Prompt Injection] (LOW): The skill retrieves payloads from a Qdrant database, creating a surface for indirect prompt injection where stored data could potentially manipulate agent behavior.\n
- Ingestion points: Data retrieved via
curlfrom Qdrant REST API endpoints (e.g.,/collections/my_collection/points/query) inSKILL.md.\n - Boundary markers: Absent; the skill does not instruct the agent to use specific delimiters or ignore instructions within retrieved payloads in
SKILL.md.\n - Capability inventory: Execution of shell commands via
bash -cand network requests viacurlinSKILL.md.\n - Sanitization: Absent; no validation or escaping of the retrieved data is mentioned in
SKILL.md.\n- [Command Execution] (SAFE): The skill usesbashandcurlfor its primary purpose of interacting with the Qdrant API. Commands are constructed using environment variablesQDRANT_URLandQDRANT_API_KEY, which is a standard and expected practice for this type of utility.
Audit Metadata