The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's workflow involves the agent reading and analyzing project source files ("Identify changed files", "Run the 7-category checklist... on each file"). This ingestion of potentially untrusted or externally-influenced code represents an attack surface for indirect prompt injection, where malicious instructions hidden in code or comments could attempt to hijack the agent's logic during verification.
Ingestion points: Project source code files identified via git or manual selection (SKILL.md).
Boundary markers: Absent. There are no instructions for the agent to use delimiters or specific safety markers to distinguish between code and embedded instructions.
Capability inventory: The agent has the capability to execute shell commands such as npm run build, npx eslint, npx vitest, and npx madge (SKILL.md).
Sanitization: Absent. The skill does not provide mechanisms for sanitizing or validating the content of the files before processing.
[COMMAND_EXECUTION]: The skill provides explicit instructions for the agent to execute shell commands to perform code verification tasks.
Evidence: The "Automated Verification Script" and "Quick Commands" sections instruct the agent to run npm run build, npx eslint, npx vitest, and npx madge (SKILL.md).

code-verification