context-optimizer
Fail
Audited by Gen Agent Trust Hub on Apr 7, 2026
Risk Level: HIGHCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides patterns for executing shell commands and Python code via the
ctx_executeandctx_execute_filetools to handle large outputs. - [EXTERNAL_DOWNLOADS]: The skill fetches documentation from well-known services including x.com and stripe.com for indexing and local searching.
- [DATA_EXFILTRATION]: The skill includes examples for reading and analyzing system log files (e.g., /var/log/app/api.log) using specialized tools. No evidence of unauthorized transmission was found.
- [PROMPT_INJECTION]: The skill implements an indirect prompt injection surface by ingesting untrusted data. Ingestion points: ctx_execute, ctx_execute_file, and ctx_fetch_and_index in SKILL.md. Boundary markers: Absent. Capability inventory: ctx_execute (shell) and ctx_execute_file (python). Sanitization: Absent.
Recommendations
- HIGH: Downloads and executes remote code from: https://docs.x.com - DO NOT USE without thorough review
Audit Metadata