internal-comms

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill operates entirely within the local environment using Python's standard library. It does not make any network requests or access sensitive system files.
  • [COMMAND_EXECUTION]: The skill provides Python scripts that the user or agent can run. These scripts are deterministic and limited to processing JSON input into Markdown or JSON output without side effects.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied data from a brief template to generate communication artifacts.
  • Ingestion points: Data enters the scripts via JSON inputs in change_announcement_builder.py, comms_calendar_builder.py, and comms_template_filler.py.
  • Boundary markers: The scripts treat inputs as data fields without explicit prompt delimiters, but the resulting output is text intended for human review.
  • Capability inventory: Analysis of all scripts confirms no subprocess calls, network activity, or file-writing capabilities exist.
  • Sanitization: Inputs are parsed via standard JSON libraries; no further sanitization is performed as the tool generates static text artifacts.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 04:56 PM
Security Audit — agent-trust-hub — internal-comms