dota2-custom-game-dev
Pass
Audited by Gen Agent Trust Hub on Aug 19, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill includes
scripts/update_references.ps1, which fetches DOTA2 API reference snapshots from an external GitHub repository (https://github.com/BigCiba/vscode-dota2-tools.git). The script clones the repository to extract specific documentation and data files. - [COMMAND_EXECUTION]: The
SKILL.mdinstructions recommend running the reference update script with theExecutionPolicy Bypassflag. This bypasses the default PowerShell security settings meant to prevent unauthorized script execution. - [PROMPT_INJECTION]: The skill processes untrusted project files and API snapshots that could be used for indirect prompt injection attacks. 1. Ingestion points: Project files like
package.jsonandtsconfig.json, and API snapshot JSONs in thereferences/directory. 2. Boundary markers: None. 3. Capability inventory: The skill performs local filesystem reading and provides data to the LLM context via search and detection scripts. 4. Sanitization: Standard JSON parsing is used, but there is no specific sanitization to prevent ingested data from being interpreted as instructions by the agent.
Audit Metadata