evidence-upload

Pass

Audited by Gen Agent Trust Hub on Aug 19, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions leverage the Bash tool to execute curl and python3 commands.
  • The curl utility is used to interact with the backend REST API for presigning and publishing assets.
  • python3 is used as a secure way to generate and parse JSON payloads using the standard json and os modules, which avoids shell escaping vulnerabilities.
  • [DATA_EXFILTRATION]: The skill facilitates the transmission of local file data to an external service endpoint defined by the $PYLOT_GATEWAY_URL environment variable.
  • This transmission is the documented primary purpose of the skill, intended to store and share visual artifacts for project documentation.
  • The flow includes local file reading, S3 presigned URL usage, and API publication steps.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests local file paths and produces public URLs that are then intended for embedding in external systems like GitHub Pull Requests or Slack threads.
  • This creates an automated bridge for data to flow from the local environment to public or shared platforms.
  • [SAFE]: The skill maintains security by utilizing environment-provided authentication tokens ($PYLOT_DISPATCH_TOKEN) and gateway URLs, ensuring that no sensitive credentials or specific backend configurations are hardcoded in the skill definition.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 19, 2026, 10:02 PM
Security Audit — agent-trust-hub — evidence-upload