The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local Python scripts located in the scripts/ directory to perform specialized tasks.
zotero_searcher.py: Conducts searches in the local Zotero database via its API.
zotero_export_bibtex.py: Fetches and appends BibTeX entries from Zotero to local bibliography files.
dedupe_bibtex.py: Merges and removes duplicates from multiple .bib files using DOI and title-based heuristics.
[EXTERNAL_DOWNLOADS]: The skill uses the requests Python library to communicate with the Zotero Local API. By default, it targets http://localhost:23119/api/ to retrieve bibliographic metadata and export entries, but the endpoint is configurable.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external literature sources, creating an attack surface where adversarial content could attempt to influence agent behavior.
Ingestion points: Processes text extracted from PDF files (via the pdf skill) and metadata/notes retrieved from the Zotero API during the 'Reading and Review' mode.
Boundary markers: No explicit delimiters or instructions are used to isolate untrusted literature content from the agent's command context in the provided scripts.
Capability inventory: The agent has the ability to write to the local file system (zotero.bib), generate LaTeX citations, and execute scripts that send API keys to a configurable network endpoint.
Sanitization: There is no specific sanitization or validation of the natural language content extracted from external sources before it is used for summarization or synthesis.

literature