author-flowglad-ows

Pass

Audited by Gen Agent Trust Hub on Sep 1, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for ingesting external specification data to guide the generation of OWS configuration files.
  • Ingestion points: The skill calls the automations.get_authoring_spec tool to fetch authoritative sections of the workflow specification (SKILL.md).
  • Boundary markers: Instructions explicitly direct the agent to verify the identity (specDigest) of the tool output and treat it as authoritative over internal knowledge.
  • Capability inventory: The skill allows the agent to write OWS YAML files and companion Python scripts, and potentially publish them via connected MCP tools.
  • Sanitization: The skill relies on the trust relationship with the vendor-provided MCP tool and does not explicitly define additional sanitization for the fetched specifications.
  • [EXTERNAL_DOWNLOADS]: Multiple reference OWS files include URIs for catalogs and profiles used in automation validation.
  • Target resources are located at https://flowglad.com/ows/catalogs/ and https://flowglad.com/ows/profiles/.
  • These are official vendor resources used for identifying schemas and extension capabilities for the platform.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 1, 2026, 02:48 PM
Security Audit — agent-trust-hub — author-flowglad-ows