zudoku-guide
Warn
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: MEDIUMPRIVILEGE_ESCALATIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [PRIVILEGE_ESCALATION]: The skill includes instructions to modify system-level configuration files such as
/etc/sysctl.confusingsudoto increase themax_user_watcheslimit for file watching. - Evidence:
echo fs.inotify.max_user_watches=524288 | sudo tee -a /etc/sysctl.confandsudo sysctl -pfound inreferences/common-errors.md. - [COMMAND_EXECUTION]: The skill instructs the agent to execute a variety of shell commands to inspect the local environment, verify package installations, and search for type definitions within the
node_modulesdirectory. - Evidence: Frequent use of
ls,cat,grep,npm, andnpxacrossSKILL.mdand reference files. - [EXTERNAL_DOWNLOADS]: The skill fetches framework documentation and structured instructions from the official Zudoku website.
- Evidence: References to
https://zudoku.dev/llms.txtand various documentation paths underhttps://zudoku.dev/inreferences/remote-docs.md. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external sources, including locally provided OpenAPI specifications and remote markdown documentation, which could contain malicious instructions.
- Ingestion points: The skill reads content from
./apis/openapi.yaml, documentation files in/pages/, and fetches remote content fromhttps://zudoku.dev/llms.txt. - Boundary markers: Absent. The instructions do not define delimiters or provide specific prompts to the agent to ignore instructions embedded within the processed data.
- Capability inventory: The skill possesses the ability to execute shell commands (
ls,grep,cat,npm) and perform system-level modifications withsudo. - Sanitization: The skill lacks evidence of input validation or sanitization of the external data before it is presented to the agent or used in commands.
Audit Metadata