cohere-debug-bundle
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Generates and runs shell scripts to aggregate environment metadata, such as Node.js and Python versions, and SDK status.
- [DATA_EXFILTRATION]: Accesses log files at /var/log/app/ to gather diagnostic information. The skill includes built-in security measures to redact Bearer tokens and sensitive API key details before data is bundled locally.
- [EXTERNAL_DOWNLOADS]: Connects to official Cohere API and status endpoints (api.cohere.com, status.cohere.com) for connectivity testing. These are well-known services appropriate for the skill's diagnostic context.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external log data which could contain malicious content. 1. Ingestion point: /var/log/app/*.log (SKILL.md). 2. Boundary markers: None present. 3. Capabilities: Bash commands (grep, curl, tar) and file writing. 4. Sanitization: Employs grep filtering and sed redaction of authentication tokens to mitigate risks from untrusted log content.
Audit Metadata