future-predictor
Fail
Audited by Gen Agent Trust Hub on May 9, 2026
Risk Level: HIGHDATA_EXFILTRATIONCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill directs the agent to automatically retrieve the last 10 commands from the bash history. This source is highly sensitive as it often contains leaked credentials, API tokens, or passwords used in command-line operations.
- [DATA_EXFILTRATION]: The skill requests broad access to all "currently open files in your IDE", which creates a risk of exposing sensitive files such as .env, private keys, or internal configuration files.
- [CREDENTIALS_UNSAFE]: Targeted access to system logs like bash history and unrestricted workspace files is a high-risk pattern for harvesting hardcoded secrets or environment credentials.
- [DATA_EXFILTRATION]: The inclusion of git diffs in context aggregation can lead to the exposure of sensitive code or secrets before they are sanitized.
- [PROMPT_INJECTION]: The skill has an indirect prompt injection surface. Ingestion points: bash history, git diff, and open files (SKILL.md). Boundary markers: absent. Capability inventory: reading files and bash history (SKILL.md). Sanitization: absent.
Recommendations
- AI detected serious security threats
Audit Metadata