The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill is designed to clone the official Contentful Optimization repository from GitHub (https://github.com/contentful/optimization.git) to retrieve the latest SDK documentation, types, and implementation examples.
[REMOTE_CODE_EXECUTION]: During the synchronization process, the skill executes lifecycle scripts from the cloned repository using pnpm knowledge:check and pnpm guides:check. This execution of remote code is performed on the vendor's official codebase.
[COMMAND_EXECUTION]: The skill uses several shell commands for its operations, including pnpm for building and testing, and python3 for running local validation scripts such as quick_validate.py.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes data from an external repository to compile reference guides, creating an attack surface for indirect prompt injection.
Ingestion points: Documentation files under documentation/internal/sdk-knowledge/ and documentation/authoring/blueprints/, public exports/types, and AGENTS.md policy files from the upstream repository.
Boundary markers: The skill instructs the agent to follow rules defined in upstream AGENTS.md files, but lacks explicit technical delimiters or escaping for the content being processed.
Capability inventory: The agent has the ability to execute shell commands (pnpm, python3, npx), read/write local files, and access the network via Git.
Sanitization: There is no mention of sanitizing or validating the text content of the ingested documentation before it is incorporated into the generated references.

optimization-sdk-sync