Back to skill
Skillv1.0.0
ClawScan security
citation-finder · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignMar 9, 2026, 2:50 PM
- Verdict
- benign
- Confidence
- high
- Model
- gpt-5-mini
- Summary
- The skill's code, instructions, and requirements are coherent with an academic citation lookup/formatter: it only performs HTTP searches (APIs + web scraping), formats citations, and requests no credentials or unusual system access.
- Guidance
- This skill appears to do what it says: it searches public APIs and scrapes Chinese search pages, then formats citations. Things to consider before installing: (1) it will make outbound HTTP requests to CrossRef, Semantic Scholar, Baidu Scholar, and CNKI — avoid sending sensitive or private data as the query. (2) Web scraping (Baidu/CNKI) may be blocked, fragile, or violate site terms; expect occasional failures or missing results. (3) There's a small documentation/code mismatch: SKILL.md's confidence thresholds differ from run.py's numeric thresholds (80/50 vs 75/45); behavior you see will follow the code. (4) No credentials are requested and no unusual file/system access occurs. If you need stricter guarantees, review the code yourself (it's included) or run it in an isolated environment and verify outputs before relying on citations in formal work.
Review Dimensions
- Purpose & Capability
- okName/description (citation lookup + formatting) align with the included scripts: English searches use CrossRef and Semantic Scholar APIs; Chinese searches scrape Baidu Scholar and CNKI; formatting produces GB/T 7714, APA, and MLA. No unrelated credentials, binaries, or config paths are requested.
- Instruction Scope
- noteRuntime instructions and code stay within the stated purpose (search, rank, and format). Minor inconsistencies: SKILL.md states confidence thresholds (>80% direct, 50–80% ask to confirm) whereas run.py uses 75/45 thresholds (CONFIDENCE_HIGH=75, CONFIDENCE_MID=45). Also the skill performs web scraping of Baidu Scholar and CNKI (documented), which is expected but may be brittle or subject to anti-scraping/TOS concerns; the code does not attempt to read unrelated files or environment variables.
- Install Mechanism
- okNo install spec (instruction-only skill) and only standard Python dependencies (requests, beautifulsoup4, rapidfuzz) are required via pip. There are no downloads from untrusted URLs or archive extraction steps.
- Credentials
- okThe skill declares no required environment variables, no credentials, and no config paths. The code does perform outbound HTTP calls to public APIs and web pages (CrossRef, Semantic Scholar, xueshu.baidu.com, kns.cnki.net) which is consistent with its stated purpose.
- Persistence & Privilege
- okSkill is not forced-always (always:false) and uses normal autonomous invocation settings. It does not modify other skills or system config and does not request persistent elevated privileges.