The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute several shell commands within the project environment.
grep -rn "<image-filename>" docs/src: Used to locate image references.
cd docs && pnpm dev --port 3005: Runs a local development server for visual verification.
pnpm lint:remark: Runs documentation linting.
pnpm build: Executes a full build of the documentation site.
[INDIRECT_PROMPT_INJECTION]: The grep command template grep -rn "<image-filename>" docs/src introduces a surface for command injection if the agent blindly interpolates user-provided text into the shell command.
Ingestion points: User-provided image filenames for conversion triggers (e.g., "convert diagram image_name.png").
Boundary markers: None provided in the command template to prevent shell meta-character execution.
Capability inventory: The skill uses shell access for file searching, running a dev server, and triggering builds.
Sanitization: The instructions do not include validation or sanitization steps for the <image-filename> variable before execution.

docs-diagrams