antivibe
Pass
Audited by Gen Agent Trust Hub on Jul 31, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection by processing untrusted source code from the project environment to generate educational content.
- Ingestion points: Project source code files (e.g., TS, JS, PY, GO, RS, JAVA) read by scripts/analyze-code.sh and scripts/capture-phase.sh.
- Boundary markers: The skill lacks explicit delimiters or specific instructions to the agent to ignore embedded commands when interpolating code snippets into reports.
- Capability inventory: The skill possesses file system read access (grep, head, find) and local file write access (mkdir, cat) to generate reports in the deep-dive/ folder. It does not have network access for data exfiltration.
- Sanitization: Content from analyzed files is included in the output without filtering or escaping for potential prompt injection patterns.
- [COMMAND_EXECUTION]: The skill executes local shell scripts to perform routine code structure analysis.
- Evidence: scripts/analyze-code.sh, scripts/capture-phase.sh, and scripts/generate-deep-dive.sh use standard system utilities like grep, head, and find.
- Context: These operations are confined to the local project environment and are directly related to the skill's primary function of documenting code.
Audit Metadata