The Agent Skills Directory

[COMMAND_EXECUTION]: The skill involves the execution of multiple system utilities and custom scripts to collect debugging information, perform network requests to the vendor's API, and package local diagnostic files. This includes usage of curl, node, npm, and tar in the provided scripts.
[DYNAMIC_CONTEXT_INJECTION]: The skill uses the !command syntax to automatically run diagnostic commands when the skill is loaded. These commands check the environment (Node.js version, package status) and verify if the EXA_API_KEY environment variable is set (displaying only its length for confirmation).
[DYNAMIC_EXECUTION]: The skill uses python3 -c to execute a short inline script for processing JSON responses from the connectivity test. This is a low-risk use of dynamic execution for basic data parsing and result counting.
[INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes results from the Exa API, which is an external and potentially untrusted source of data.
Ingestion points: Data from https://api.exa.ai/search is retrieved via curl and stored in /tmp/exa-debug.json, and search contents are fetched via the exa-js library.
Boundary markers: The instructions do not define clear boundaries or provide explicit instructions for the agent to ignore commands that might be embedded within the search result data.
Capability inventory: The skill has access to powerful tools including Bash (permitting curl, tar, node, grep) and Read for handling the collected data.
Sanitization: The skill performs minimal sanitization, primarily focusing on formatting JSON output for logging purposes rather than filtering or validating the content of the search results.

exa-debug-bundle