The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches bibliographic metadata from authoritative external sources, specifically api.crossref.org (Crossref) and openlibrary.org (Open Library). These are well-known, trusted academic services. The operations are performed using the requests library with appropriate timeouts and rate-limiting.
[DATA_EXPOSURE]: The skill reads and writes to local files (e.g., reference_database.json, citation_styles.json) to manage the reference library and configuration. This file access is limited to the skill's own data directory and is consistent with its primary purpose of managing a local database of references.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external APIs (Crossref/OpenLibrary). While this presents a surface for indirect prompt injection if an attacker were to poison metadata in those repositories, the skill itself does not perform any dangerous operations (like eval or command execution) on the retrieved data. The risk is considered low and inherent to the utility's function.
[DYNAMIC_EXECUTION]: The documentation files COMPLETION_SUMMARY.md and PROJECT_COMPLETION_SUMMARY.md contain Python boilerplate and exec calls intended to generate or write the documentation files. These files are artifacts of the development process and are not part of the active skill logic used by the agent during runtime.

citation-manager