The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands to verify the presence of the EXTRUCT_API_TOKEN environment variable and provides instructions for the user to export it if missing.
[EXTERNAL_DOWNLOADS]: The skill is configured to fetch live documentation from https://www.extruct.ai/docs at runtime to ensure local reference files are up to date.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it ingests and processes untrusted data from multiple sources.
Ingestion points: Data is read from local context files (claude-code-gtm/context/) and retrieved from external API responses (e.g., company_description from search results).
Boundary markers: There are no explicit delimiters or instructions provided to the agent to ignore potentially malicious content embedded within the ingested data.
Capability inventory: The skill has the capability to perform network requests to the vendor's API and documentation site, read local configuration files, and overwrite local reference files (references/discovery-api.md, references/search-filters.md).
Sanitization: No evidence of sanitization or validation of the external content was found before it is used in subsequent logic or prompt construction.

list-building