jira-daily
Warn
Audited by Gen Agent Trust Hub on Mar 26, 2026
Risk Level: MEDIUMDATA_EXFILTRATIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [DATA_EXFILTRATION]: T he s kill a ccesses a s ensitive l ocal c onfiguration f ile.
- E vidence: I n
SKILL.md, t he s kill e xecutescat ~/.config/.jira/.config.ymlt o e xtract p roject s ettings. T his p ath t ypically c ontains J ira a uthentication t okens o r s ession d ata. - [PROMPT_INJECTION]: T he s kill h as a n i ndirect p rompt i njection s urface t hrough t he p rocessing o f e xternal d ata.
- I ngestion p oints: C ontent i s r etrieved f rom
jira issue list(summaries a nd d escriptions) a ndgit log(commit m essages) a s s een i nSKILL.md. - B oundary m arkers: A bsent. T he s kill i nstructions i nterpolate t hese d ata s ources d irectly i nto p rompts f or s ub-agents w ithout d elimiters o r i nstructions t o i gnore e mbedded c ommands.
- C apability i nventory: T he s kill u tilizes
Bash(with a ccess t o j ira, g it, a nd c at c ommands),Read, a ndTodoWritet ools a cross m ultiple f iles. - S anitization: A bsent. T here i s n o e vidence o f v alidation o r f iltering o f t he c ontent r etrieved f rom J ira o r G it b efore i t i s p rocessed b y t he A I.
- [EXTERNAL_DOWNLOADS]: T he s kill d ocumentation r eferences a n e xternal t ool.
- E vidence:
SKILL.mdi dentifies a d ependency o n t hejira-clit ool, p roviding a l ink t o t he o fficial G itHub r epository o f t he p roject.
Audit Metadata