booth-beat
Pass
Audited by Gen Agent Trust Hub on Apr 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute the
booth lsandbooth reportscommands to retrieve the current state of the workspace. - [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting untrusted data from
.claude/progress.mdandbooth reports. It specifically instructs the agent to perform autonomous actions ("Spin immediately. Don't ask user.") based on the content of these files, which could lead to unintended task execution if the files are compromised. - Ingestion points:
.claude/progress.mdand outputs frombooth reports(SKILL.md) - Boundary markers: None present; the agent is instructed to act directly on the findings.
- Capability inventory: Execution of
boothCLI tools and the ability to "spin" new tasks/decks autonomously. - Sanitization: No sanitization or validation of the input data is mentioned before it is processed.
Audit Metadata