security-scanning-security-sast
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes standard security command-line tools such as
bandit,semgrep,eslint, andgosecto perform analysis. - The Python implementation in
SKILL.mdusessubprocess.runwith list-based arguments and hardcoded configuration paths to invoke these tools safely. - [EXTERNAL_DOWNLOADS]: The skill provides instructions to install security tools from official package registries.
- Fetches Python packages like
banditandsemgrepfrom PyPI. - Installs Node.js security plugins from the npm registry.
- Downloads scanning rules from Semgrep's official registry using the
--config=autoand--config=p/*flags. - [INDIRECT_PROMPT_INJECTION]: As a security scanning tool, this skill is designed to ingest and process untrusted source code as its primary input.
- Ingestion points: Reads source files (e.g.,
.py,.js,.java) and configuration files (e.g.,package.json,requirements.txt) from the project directory for analysis. - Boundary markers: Does not implement specific delimiters for the agent to distinguish between code content and instructions, relying on the underlying LLM's capability to treat the data as code for analysis.
- Capability inventory: Includes the ability to execute shell commands (SAST tools) and read files throughout the project path.
- Sanitization: The Python scanner uses
json.loads()to safely parse the output of security tools before summarizing results.
Audit Metadata