incident-retrospective

Pass

Audited by Gen Agent Trust Hub on Jul 2, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it processes external data from Slack and Linear to draft its output.
  • Ingestion points: Slack messages and Linear issues are retrieved and processed (SKILL.md).
  • Boundary markers: No explicit delimiters or safety instructions are defined in the instructions for the gathered content.
  • Capability inventory: The agent can write to Notion and perform HTTP requests via curl.
  • Sanitization: No sanitization of the external data is specified before it is used to generate the retrospective draft.
  • [COMMAND_EXECUTION]: The skill uses curl to register its automation with the OpenHands backend. This operation uses platform-provided environment variables (OPENHANDS_HOST, OPENHANDS_AUTOMATION_API_KEY) and targets vendor infrastructure, which is standard and safe behavior for a skill from this author.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 2, 2026, 11:05 AM
Security Audit — agent-trust-hub — incident-retrospective