The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes scripts/crawlora.sh, which uses the system curl command to communicate with the API. The script is properly quoted to prevent shell injection from input arguments.
[EXTERNAL_DOWNLOADS]: Fetches data from api.crawlora.net, which is the official API endpoint for the skill's authoring organization, crawlora-org.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted content from public Upwork listings. While this content is user-generated and could contain malicious instructions, the skill does not execute the data or use it in sensitive operations.
Ingestion points: Data retrieved from /upwork/search, /upwork/job/{id}, and /upwork/freelancer/{id}.
Boundary markers: None present.
Capability inventory: Network read capabilities via curl in scripts/crawlora.sh.
Sanitization: None applied to the fetched Upwork data.

upwork-research