deliver

Pass

Audited by Gen Agent Trust Hub on Jul 29, 2026

Risk Level: SAFECREDENTIALS_UNSAFEDATA_EXFILTRATION
Full Analysis
  • Credential Access: The skill accesses the file path /credentials/webhook.env to retrieve webhook URLs. This is the intended location for secret storage within the agent workspace.
  • Network Operations: The skill transmits briefing data to external chat services using HTTP POST requests. The destination is determined by the configuration in the credentials file.
  • Data Ingestion Point: The script scripts/deliver.py reads and processes markdown files from the workspace. This represents a point where untrusted data enters the skill's context for summarization.
  • Sanitization and Boundaries: The skill uses regex for text formatting but does not employ specific boundary markers or sanitization to differentiate between data and potential instructions during the condensation process.
  • Capability Inventory: The skill possesses network communication capabilities via urllib.request.urlopen to perform its delivery task.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 29, 2026, 01:46 PM
Security Audit — agent-trust-hub — deliver