bd-tracker-updater

Pass

Audited by Gen Agent Trust Hub on May 9, 2026

Risk Level: SAFE
Full Analysis
  • [DATA_EXFILTRATION]: The skill performs network operations via HTTP POST to an endpoint configured through environment variables ($BD_Tracker_Base_URL and $BD_Tracker_Endpoint). This ensures that the destination is managed by the user/administrator. No sensitive local files or hardcoded credentials are accessed or transmitted.
  • [COMMAND_EXECUTION]: No shell commands or system-level executions are present in the skill or its sub-components.
  • [REMOTE_CODE_EXECUTION]: There are no patterns of downloading or executing remote scripts, binary executables, or unverifiable packages.
  • [PROMPT_INJECTION]: The skill contains no instructions attempting to bypass safety guidelines or extract system prompts. It was analyzed for indirect injection surfaces due to its data processing nature:
  • Ingestion points: Data is ingested from the conversation context, such as research reports (Step 2A of org-creation/SKILL.md).
  • Boundary markers: While specific delimiters are not explicitly defined for the input, the skill uses structured extraction and mapping which limits the risk of interpreting data as instructions.
  • Capability inventory: The skill utilizes an HTTP POST method for data submission to a configured endpoint.
  • Sanitization: Strict validation rules, domain normalization, and a mandatory human-in-the-loop confirmation step (Step 3) ensure that data is reviewed and verified before submission, effectively mitigating indirect injection risks.
Audit Metadata
Risk Level
SAFE
Analyzed
May 9, 2026, 12:34 PM