The Agent Skills Directory

[DATA_EXFILTRATION]: The CLI tool supports an --deliver webhook:<url> flag which allows sending command output (potentially containing telemetry, camera URLs, or state information) to an external POST endpoint.
Evidence: 'webhook: | POST the output body to the URL (application/json or application/x-ndjson when --compact)'
[INDIRECT_PROMPT_INJECTION]: The 'automatic learning' system (recall and teach commands) processes user queries and stores them to influence future agent behavior via 'playbooks' and 'notes'.
Ingestion points: User queries passed to juneoven-pp-cli recall and stored via juneoven-pp-cli teach.
Boundary markers: The instructions provide a protocol for the agent to follow but do not define strict security boundaries for the learned content.
Capability inventory: The skill has access to network operations (API/Webhooks), local file writes (SQLite database, configuration files), and shell command execution.
Sanitization: Instructions mention a 'PII rule' for the agent to manually strip identifiers, but no automated sanitization is described.
[DYNAMIC_EXECUTION]: The skill uses a 'playbook' system that dynamically reconstructs and executes sequences of shell commands stored in a local SQLite database.
Evidence: 'replay Playbook.steps in order, substituting Playbook.slots_resolved entries for the entity slot tokens.'
[EXTERNAL_DOWNLOADS]: The skill installs binary tools and dependencies from external repositories.
Evidence: go install github.com/mvanhorn/printing-press-library/library/devices/juneoven/cmd/juneoven-pp-cli@latest
Evidence: npx -y @mvanhorn/printing-press-library install juneoven --cli-only
[COMMAND_EXECUTION]: The skill's primary function is to execute shell commands via the juneoven-pp-cli binary to control hardware and manage local state.

pp-juneoven