wiki-term-research
Pass
Audited by Gen Agent Trust Hub on Apr 15, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exh i b i t s a n att ac k su r fac e f o r in d i rec t prom p t in jec t i o n due to it s rel i an c e on ext ern al data.
- In ges t i o n p o in t s: In SKIL L.m d an d researc h-guide.m d, the agen t is in struc t ed to use WebFetc h to ret r i e v e con t en t from au t h o r i t a t i v e in d u st ry sou r c es.
- Bou n d ary mark ers: The re are no spec i f ied del im i t ers or in struc t i o n s to t rea t fetc hed con t en t as un t r u s t ed data, in c rea s i n g the ris k th a t the agen t m i gh t fol l o w in st ruc t i o n s em bed ded wit h in th a t con t en t.
- Cap a b i l i t y in v en t ory: The agen t can rea d loc a l wik i d i rec t or ies an d exec u t e web searc h/fetc h to ol s, wh ic h cou l d be m i s u s ed in a m u l t i-st ep att ac k ch a i n.
- San i t i z a t i o n: The work f l o w doe s n o t in c l u d e st ep s f o r san i t i z i n g or v a l i d a t i n g fetc hed con t en t bef o r e proc es s i n g it f o r wik i en t ry com p o s i t i o n.
Audit Metadata