The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user or agent to install the keenable-pp-cli tool using vendor-specific resources including npx -y @mvanhorn/printing-press-library and go install github.com/mvanhorn/printing-press-library/library/ai/keenable/cmd/keenable-pp-cli@latest.
[DATA_EXFILTRATION]: The CLI tool described in the skill includes a --deliver webhook:<url> flag. This feature allows the agent to automatically POST command results, which may contain sensitive research data or metadata, to an external URL.
[INDIRECT_PROMPT_INJECTION]: The skill fetches untrusted markdown and web content via the fetch and web-search commands. Because this content is then processed by the agent and can be 'learned' via the teach and playbook amend commands, it creates a risk where malicious instructions embedded in web pages could be persisted in the local store and later executed as 'ground truth' steps or notes.
Ingestion points: Untrusted web content is ingested through the keenable-pp-cli fetch and keenable-pp-cli web-search commands.
Boundary markers: While the agent is instructed to use JSON output (--agent), there are no specific markers or sanitization steps provided to prevent the agent from following instructions embedded within the fetched research data.
Capability inventory: The agent has Read Bash permissions and can execute the keenable-pp-cli binary, which has capabilities to write to local storage and perform network requests.
Sanitization: The skill provides a 'PII rule' to strip identifiers from queries, but it does not specify sanitization for the content of the research notes or playbooks being stored.
[COMMAND_EXECUTION]: The skill's primary function is to drive the keenable-pp-cli binary through a series of shell commands, including complex research and learning management tasks.

pp-keenable