The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes meeting transcripts fetched from an external API, which provides a surface for indirect instructions to be processed by the agent.
Ingestion points: Transcripts are ingested via the sync and transcripts pull commands as described in SKILL.md.
Boundary markers: The instructions lack explicit boundary markers or warnings for the agent to disregard instructions found within the transcripts.
Capability inventory: The skill can execute local CLI commands, write to the filesystem (transcripts export), and perform network operations (sync, webhook delivery).
Sanitization: No sanitization or filtering of the transcript content is mentioned before it is processed.
[EXTERNAL_DOWNLOADS]: The skill instructs the user or agent to install binary tools from external repositories.
Evidence: Installation commands include npx -y @mvanhorn/printing-press-library and go install github.com/mvanhorn/printing-press-library/... (SKILL.md).
Context: These resources originate from the vendor associated with the skill.
[COMMAND_EXECUTION]: The skill facilitates the execution of the fireflies-pp-cli binary with arguments that can be influenced by external data or user input.
[DATA_EXFILTRATION]: The integrated CLI tool includes a feature that can route command output to external URLs.
Evidence: The --deliver webhook:<url> flag enables POSTing data to an arbitrary remote endpoint (SKILL.md).
Context: This is a documented feature for automation but represents a functional path for data exfiltration if misused.

pp-fireflies