The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides templates for curl and PowerShell commands to interact with identity provider endpoints. These commands are limited to metadata retrieval and account existence checks, which are consistent with authorized reconnaissance activities.
[EXTERNAL_DOWNLOADS]: The skill interacts with the public APIs of well-known services, including Microsoft Entra (Azure AD), Okta, and Google Workspace. These references are used for fingerprinting and metadata discovery and do not involve the execution of remote scripts or unverified code.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted inputs such as domains and employee names for its discovery workflows. 1. Ingestion points: Target domains and harvested employee names (SKILL.md §7, §11, §12). 2. Boundary markers: Absent; the skill does not use specific delimiters to isolate external data. 3. Capability inventory: Execution of network requests via curl and PowerShell. 4. Sanitization: The skill relies on standard variable interpolation within command templates. This surface is assessed as safe because the skill is constrained to reconnaissance of official provider endpoints and lacks dangerous execution capabilities.
[SAFE]: No malicious code, obfuscation, or data exfiltration mechanisms were detected. The skill follows mature security practices, including identifying its own detectability level and providing a clear scope boundary between enumeration and authenticated testing.

identity-provider-recon