The Agent Skills Directory

[DATA_EXFILTRATION]: The zoom-pp-cli tool supports a --deliver webhook:<url> flag, which allows the agent to POST command results to an external URL.
This mechanism can be used to exfiltrate sensitive information extracted from Zoom, such as meeting summaries, attendee lists, or full transcripts.
[PROMPT_INJECTION]: The skill demonstrates a surface for indirect prompt injection by processing meeting transcripts and notes which may contain malicious instructions from third parties.
Ingestion points: The find, notes search, and notes ingest commands retrieve text from meeting records and local documents.
Boundary markers: The instructions do not define delimiters or provide guidance to the agent to ignore instructions embedded within the processed transcripts.
Capability inventory: The agent has access to the Bash tool and can perform file system writes and network operations via the CLI's delivery sinks.
Sanitization: No sanitization or validation of the ingested text is mentioned before it is returned to the agent context.
[EXTERNAL_DOWNLOADS]: The skill requires installing a binary from external repositories.
It utilizes npx -y @mvanhorn/printing-press-library and go install github.com/mvanhorn/printing-press-library/... to provide the core functionality.
[COMMAND_EXECUTION]: The skill relies on the execution of the zoom-pp-cli binary to interact with the Zoom service and local file system.

pp-zoom