The Agent Skills Directory

[DYNAMIC_EXECUTION]: The helper library scripts/check-lib.mjs utilizes the import() function to dynamically load Node.js modules from computed file paths. This is used by the importPackage function to load pinned dependencies required for the skill's auditing tasks.
[COMMAND_EXECUTION]: The skill's verification and self-testing framework in scripts/check-lib.mjs and scripts/selftest.mjs uses spawnSync to execute Node.js scripts. This is a functional requirement for running and testing the auditing logic against fixtures.
[INDIRECT_PROMPT_INJECTION]: The auditing scripts check-error-spine.mjs and check-error-net.mjs perform automated scans of the user's source code and configuration files. This creates a surface for indirect prompt injection where malicious content within a project being analyzed could attempt to influence the AI agent's behavior.
Ingestion points: Project source files (.ts, .tsx, .json) are read into memory using readFileSync in the analysis scripts.
Boundary markers: Absent; the tools perform regex-based heuristic scans on raw file content.
Capability inventory: The skill possesses capabilities to write files (writeFileSync), execute subprocesses (spawnSync), and load external code (import()).
Sanitization: Absent; analyzed code snippets may be included in problem reports generated by the tool.

error-handling-typed-results