agentic-security-scanner
Warn
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: MEDIUMREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONMETADATA_POISONING
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill facilitates the download and execution of software from external, unverified sources. It instructs the use of
npx @clear-capabilities/agentic-security-scanner, which executes code directly from the NPM registry, and installs a plugin fromhttps://github.com/Clear-Capabilities/agentic-security. - [METADATA_POISONING]: There is a discrepancy between the declared author ('reason-machines') and the attribution within the skill content ('ara.so' and 'Clear-Capabilities'). This inconsistency can be misleading regarding the tool's origin.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted code and configuration files, creating a surface where malicious instructions could influence the agent's behavior. 1. Ingestion points: The scanner reads all project source code and configuration files like
.agentic-security/config.json. 2. Boundary markers: No explicit delimiters are used to separate untrusted file content from agent instructions. 3. Capability inventory: The skill can modify files, execute shell commands, and make network requests via webhooks. 4. Sanitization: No sanitization of analyzed file content is specified. - [EXTERNAL_DOWNLOADS]: The skill relies on external, unverified repositories and package registries for its operation.
- [COMMAND_EXECUTION]: The skill uses shell commands and
npxto perform its primary scanning and remediation functions, executing logic on the host system.
Audit Metadata