Zotero

**Judgement:** This skill is a mature, feature-rich interface to Zotero libraries, well-suited for serious academic workflows, automation, and library hygiene tasks. It offers broad functionality around item management, metadata enrichment, and PDF handling with careful attention to rate limits and duplicate detection. **Key strengths:** - Comprehensive Zotero coverage: listing/searching, adding by DOI/ISBN/PMID, updating metadata/tags, deletion, collections/tags management, and multiple export formats. - Advanced automation for library quality: DOI completion (CrossRef), citation cross-referencing, batch-add, PDF presence checks, and OA PDF fetching. - Sensible safety defaults: dry-run for DOI-filling and PDF fetching, trash-first deletion, explicit confirmation flags for destructive actions, and constrained metadata updates. - Good integration surface: JSON output mode and no external dependencies make it straightforward to wire into OpenClaw pipelines. **Main risks / limitations:** - **Data modification risk:** Commands like `update`, `delete` (especially `--permanent`), `find-dois --apply`, and `fetch-pdfs` can irreversibly change the user’s library or attachments if misused or scripted incorrectly. - **External service dependency:** Functionality such as DOI resolution and OA PDF fetching depends on Zotero’s translation server, CrossRef, Unpaywall, Semantic Scholar, and DOI endpoints; outages or API changes can cause failures or degraded accuracy. - **Performance on large libraries:** Operations that scan many items (`check-pdfs`, `fetch-pdfs`, some batch operations) can be slow due to deliberate rate limiting and library size; they should be scoped with `--limit` or `--collection` in large setups. **Recommended scenarios:** - Managing or curating a user’s Zotero library when they have provided valid Zotero API credentials and explicitly want automated assistance. - Automating bibliography creation and export (BibTeX/RIS/CSL-JSON) from specific collections or tag filters for use in manuscripts or reports. - Cleaning up and enriching research libraries by filling missing DOIs, checking for missing PDFs, and attaching OA PDFs where legally available. - Batch-importing reading lists or citation dumps (by DOIs/ISBNs/PMIDs) while minimizing duplicates and adding consistent tags for organization. **Caution:** When the user’s intent is ambiguous, avoid destructive or high-impact commands (`delete --permanent`, `find-dois --apply`, unscoped `fetch-pdfs`) and prefer read-only operations (`items`, `search`, `get`, `export`, dry-run modes) or ask for explicit confirmation and scoping (collection/limit).