The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it is designed to ingest and process untrusted data from a user's codebase (source files, configuration manifests, and environment files) across its scanning modules.\n
Ingestion points: Codebase files and project configurations processed by the security-scan, config-scan, secrets-scan, and dependency-scan modules.\n
Boundary markers: Absent. The instructions do not specify the use of delimiters or 'ignore' instructions when reading content from external files.\n
Capability inventory: The skill possesses capabilities to read/write files and execute shell commands (e.g., npm audit, pip-audit, /security-scan --fix) which could be abused if an injected instruction is obeyed.\n
Sanitization: Absent. Content from scanned files is processed via regex pattern matching without explicit sanitization or validation of the input strings before they enter the agent context.\n- [COMMAND_EXECUTION]: The skill utilizes several security-related command-line tools, including npm audit, pip-audit, safety, bundler-audit, and git, to perform project discovery and vulnerability analysis. This behavior is consistent with the skill's primary purpose as a security scanner.

security