watch-and-act
Pass
Audited by Gen Agent Trust Hub on Aug 11, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: An indirect prompt injection surface is present because the skill monitors and processes data from external, untrusted sources such as CI logs or health check responses.
- Ingestion points: External state queried via tools like gh or curl in SKILL.md.
- Boundary markers: Absent; the instructions do not specify the use of delimiters when processing external signal content.
- Capability inventory: The skill uses Bash, Write, ScheduleWakeup, and Agent tools.
- Sanitization: Absent; the skill records and acts on external evidence without specific filtering mentioned in the instructions.
- [COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute shell commands for querying the state of external resources.
- [EXTERNAL_DOWNLOADS]: The skill performs network operations (e.g., curl) to fetch status information from external endpoints as part of its primary polling functionality.
Audit Metadata