The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill workflow includes the command npx ai-devkit@latest, which fetches and executes code from the npm registry at runtime. Running unversioned packages introduces a dependency on the security of the public registry and the package maintainer.
[COMMAND_EXECUTION]: The skill executes shell commands using npx to perform memory searches and validate entry points during the documentation process.
[EXTERNAL_DOWNLOADS]: Generated HTML documentation includes a reference to https://cdn.jsdelivr.net/npm/mermaid/dist/mermaid.min.js, which is an external script dependency from a well-known CDN used to render diagrams.
[PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection because it processes arbitrary code provided by users.
Ingestion points: The skill reads file content, function signatures, and directory structures from user-specified entry points in the Collect Source Context step.
Boundary markers: No specific delimiters or instructions to ignore embedded commands are used when the agent processes the source code content.
Capability inventory: The agent has the capability to execute shell commands (npx) and write files to the local environment (docs/ai/implementation/).
Sanitization: The skill does not validate or sanitize the ingested source code for malicious instructions before analysis.

document-code