The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process potentially untrusted content from source code files, dependency manifests, and infrastructure-as-code templates. This creates a surface for indirect prompt injection where malicious instructions hidden in the scanned files could attempt to manipulate the agent's logic during the reporting or blocking phases.
Ingestion points: The skill reads file content using the Read and Grep tools and processes entire directories through mcp_snyk_snyk_code_scan, mcp_snyk_snyk_sca_scan, and mcp_snyk_snyk_iac_scan tools.
Boundary markers: There are no explicit delimiters or specific 'ignore embedded instructions' directives defined to separate the data being scanned from the agent's operating instructions.
Capability inventory: The skill has access to Bash for command execution and several Snyk-specific MCP tools for scanning and providing feedback.
Sanitization: While the skill provides structured reporting and filters findings based on Git diffs, it does not explicitly sanitize the textual descriptions of vulnerabilities returned by the tools before presenting them to the agent or user.
[COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute git diff and git status commands to identify changed files. These are standard development operations used here to scope the security scans effectively.

secure-at-inception