k-skill-setup
Warn
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPERSISTENCEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches and executes the
@nomadamas/k-skillandskillspackages from the npm registry usingnpxcommands during the setup process. - [REMOTE_CODE_EXECUTION]: The skill directs the agent to execute a remote CLI command (
npx -y @nomadamas/k-skill@0 instruct k-skill-setup) and adopt the resulting output as the primary set of instructions, which effectively executes dynamic logic retrieved at runtime. - [PERSISTENCE]: The skill provides automated scripts to configure persistent execution on the host system, utilizing
crontabon Unix-based systems andschtaskson Windows to perform daily update checks. - [DYNAMIC_EXECUTION]: The skill creates executable shell scripts (
check-skill-updates.sh) and Windows batch files (check-skill-updates.cmd) at runtime to facilitate the scheduled tasks. - [INDIRECT_PROMPT_INJECTION]:
- Ingestion points: The skill relies on external instruction content fetched via the
@nomadamas/k-skillCLI tool during execution. - Boundary markers: The skill lacks explicit delimiters or warnings to the agent to treat the external CLI output as untrusted or to ignore embedded instructions.
- Capability inventory: The skill utilizes shell execution (
npx), file creation (cat >), permission modification (chmod), and system scheduling (crontab,schtasks) across its scripts. - Sanitization: There is no evidence of validation or sanitization of the dynamic instructions fetched from the remote source before they are processed by the agent.
Audit Metadata