The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses several command-line tools including curl for API interactions, git for update checks, and text processing utilities like jq, awk, sed, sort, and tail to handle API responses and version tags.
[EXTERNAL_DOWNLOADS]: The skill interacts with the Zenodo API (both production and sandbox environments) to upload research artifacts and metadata. It also connects to GitHub to fetch version tags for the update check routine.
[REMOTE_CODE_EXECUTION]: The skill implements a self-update mechanism that checks for newer tags in its origin Git repository. While it requires explicit user consent to perform a git pull, this pattern allows for the remote modification of the skill's executable instructions and scripts.
[INDIRECT_PROMPT_INJECTION]: The skill acts as a bridge between the user/external data and the Zenodo API. It ingests research metadata such as titles, descriptions (which allow HTML), and creator information. If an agent is instructed to deposit data from an untrusted source, these fields could be used to store malicious instructions, although the risk is limited by Zenodo's schema validation and the intended research context.

zenodo-skill