The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it retrieves book titles and descriptions from the external OpenLibrary API and processes them. Maliciously crafted data in the API response could attempt to influence the agent's behavior when displayed or stored in memory.\n
Ingestion points: Data fetched from the OpenLibrary API in src/openlibrary_client.py.\n
Boundary markers: No specific delimiters or boundary markers are used when presenting external metadata to the agent.\n
Capability inventory: The skill can write discoveries to the agent's memory using store_discoveries in src/cli.py.\n
Sanitization: Descriptions are truncated to 500 characters and subjects are limited to 20 in src/openlibrary_client.py, which provides a basic constraint on payload size.\n- [EXTERNAL_DOWNLOADS]: The skill fetches bibliographic data from openlibrary.org, a well-known and trusted public service. It also references the uv installer from astral.sh.\n- [REMOTE_CODE_EXECUTION]: The run.sh script includes a command to install the uv package manager via curl | sh. This is a remote code execution pattern, although it targets a well-known and reputable developer tool.\n- [COMMAND_EXECUTION]: The skill uses uv run to execute local Python scripts and sanity checks to facilitate book discovery functionality.

discover-books