The Agent Skills Directory

[Command Execution]: The skill executes external commands via the Azure CLI (az) to manage cloud resources and spawns local worker processes to perform document assessment. This is a potential consideration for command safety, but the skill implements hardening measures such as isolated execution flags and OS-level resource limits (Windows Job Objects and Linux rlimits) to bound the worker's capabilities.
[Indirect Prompt Injection Surface]: The skill parses untrusted documents like PDFs and Office files to extract structural metadata. While parsing external files is a security consideration, the skill uses defensive libraries like defusedxml and Python audit hooks to deny unauthorized network or filesystem access during parsing.
[Sensitive Data Redaction]: Logic within the skill proactively identifies and redacts secrets such as API keys, tokens, and connection strings from inputs and outputs. This ensures that any credentials used for Azure service integration are not accidentally exposed in results or logs.
[External Dependencies]: The skill specifies well-known Python packages for document structural analysis. These dependencies are standard for the skill's intended purpose and are utilized in a restricted, read-only manner to extract metadata.

foundry-iq