The Agent Skills Directory

Command Execution via Local Scripts: The skill executes a bundled helper script (launch_codex_security_mcp) to manage scan scopes and inventories. This is a core component of its functionality, used to bridge the agent with specialized security tools.
Evidence: SKILL.md and references/scan-artifacts-and-ledger.md describe several shell commands that invoke the script located in the <plugin_dir>/scripts/ directory.
Indirect Prompt Injection Surface: The skill's primary purpose is to ingest and analyze untrusted source code from external repositories. This data processing creates a surface where malicious instructions embedded in the scanned files (e.g., in comments or documentation) could attempt to influence the agent's reasoning.
Ingestion points: The skill resolves and reads the content of files within the target repository scope (documented in SKILL.md, Workflow Step 1).
Boundary markers: There are no explicit instructions or delimiters defined to isolate the source code content from the agent's operational instructions.
Capability inventory: The skill has the capability to execute shell commands via local scripts, write scan results to JSON files, and finalize reports (documented in SKILL.md, Steps 1 and 4).
Sanitization: The instructions do not mention specific sanitization or filtering of the source code before it is analyzed by the agent.

security-scan