The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of the ankiweb-pp-cli tool via npx or go install from the vendor's repository. \n
npx -y @mvanhorn/printing-press-library install ankiweb --cli-only \n
go install github.com/mvanhorn/printing-press-library/library/education/ankiweb/cmd/ankiweb-pp-cli@latest \n- [CREDENTIALS_UNSAFE]: The skill provides mechanisms to access sensitive session cookies from the Chrome browser or environment variables to authenticate with the AnkiWeb service. \n
auth login --chrome \n
ANKIWEB_COOKIES environment variable \n- [DATA_EXFILTRATION]: The tool supports a --deliver webhook:<url> feature that allows the user or an agent to POST command results, which may include sensitive deck information, to an arbitrary external URL. It also includes a feedback mechanism that can send local data to an external endpoint if configured. \n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted content from the AnkiWeb shared deck catalog, such as deck titles, descriptions, and user reviews. \n
Ingestion points: Data enters the context via shared search, shared info, and decks command outputs. \n
Boundary markers: While output is structured in a JSON provenance envelope, there are no specific delimiters or instructions for the agent to ignore potentially malicious content within deck descriptions. \n
Capability inventory: The agent has the Read Bash tool and the ability to write output to local files or remote webhooks via the CLI flags. \n
Sanitization: No explicit sanitization or filtering of the retrieved AnkiWeb metadata is performed before it is presented to the agent.

pp-ankiweb