flyio-webhooks-events
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill monitors application logs via the
fly logscommand, which allows untrusted data from the application's standard output to enter the agent's context.\n - Ingestion points: Application logs are streamed into the environment through the
fly logscommand inSKILL.md.\n - Boundary markers: The skill contains instructions for the agent to exclude log bodies and user data from receipts, but it does not implement technical delimiters or boundary markers for the ingested data.\n
- Capability inventory: The skill allows the agent to write files and execute network commands via
Bash(curl).\n - Sanitization: The instructions recommend redaction of sensitive incident evidence and using synthetic fixtures for testing.\n- [DATA_EXFILTRATION]: The skill facilitates deployment notifications to Slack.\n
- Evidence: Step 4 in
SKILL.mdusescurlto send application status data to a Slack webhook URL.\n - Context: The network operation targets a well-known service (Slack) and is a core part of the skill's deployment automation functionality.
Audit Metadata