The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes various local scripts and shell commands to manage the deployment process, including bun src/admin/make-layer-public.ts, aws sts get-caller-identity, and a Docker verification script ./run.sh.
[CREDENTIALS_UNSAFE]: The agent is instructed to export AWS credentials into the active shell environment using eval "$(aws configure export-credentials --format env)". This makes sensitive access keys available to the current session to facilitate Lambda operations.
[EXTERNAL_DOWNLOADS]: The skill fetches version metadata from Microsoft's official Playwright GitHub repository and performs network checks via curl to verify binary availability on the vendor's domain (remotion.media).
[INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources and script outputs which are then used to modify local source code.
Ingestion points: Reads configuration data from https://github.com/microsoft/playwright/blob/main/packages/playwright-core/browsers.json and parses the stdout output from the make-layer-public.ts script (SKILL.md).
Boundary markers: None identified; the instructions do not specify the use of delimiters or security warnings when parsing external data.
Capability inventory: The skill maintains capabilities to execute shell commands (bun, bash), perform file writes across multiple project packages, and execute network operations (curl).
Sanitization: No specific sanitization or validation steps are mentioned for the data ingested from external URLs or command outputs before they are interpolated into source files.

update-chrome-binaries-test-region