ai-listening
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFECREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [CREDENTIALS_UNSAFE]: The skill checks for an API key stored in a hidden local directory (
~/.aisa/key) or an environment variable to authenticate with the AIsa service. - [EXTERNAL_DOWNLOADS]: Clones the
new-term-radartool from the author's (zhuyansen) GitHub repository if it is not already present locally. - [COMMAND_EXECUTION]: Executes shell commands to clone the repository and run the Python-based radar utility (
python -m radar) with various flags. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external platforms including X (Twitter), GitHub, and website sitemaps, which could contain malicious payloads intended to influence the agent.
- Ingestion points: Data is read from the
listening.jsonfile produced by the radar tool (file referenced in SKILL.md). - Boundary markers: The instructions do not define clear delimiters or include warnings to ignore instructions embedded in the collected data.
- Capability inventory: The skill has the ability to execute shell commands and perform file read/write operations (defined in SKILL.md).
- Sanitization: No sanitization, filtering, or validation steps are specified for the external content before it is processed into the final report.
Audit Metadata