overnight-runner

Pass

Audited by Gen Agent Trust Hub on Apr 1, 2026

Risk Level: SAFE
Full Analysis
  • [DATA_EXFILTRATION]: The skill includes a pattern for sending status notifications to Telegram via its official API (api.telegram.org). This uses environment variables for the bot token and chat ID, adhering to secure credential management practices.
  • [COMMAND_EXECUTION]: Provides bash scripts to wrap the agent's execution, facilitating retries, logging, and status monitoring. These scripts utilize standard system utilities such as jq, curl, and tee for operational automation.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external JSON files, which represents a surface for indirect prompt injection if the task queue contains untrusted content.
  • Ingestion points: Reads task definitions and status updates from tasks/batch-queue.json and tasks/checkpoint.json (SKILL.md).
  • Boundary markers: None explicitly defined in the example prompt templates.
  • Capability inventory: The skill facilitates file system interaction and network notification via shell scripts.
  • Sanitization: No specific sanitization or validation of the JSON content is described in the prompt patterns.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 1, 2026, 05:34 AM
Security Audit — agent-trust-hub — overnight-runner