The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx -y @nomadamas/k-skill@0 to download and execute a command-line tool for fetching instructions and managing updates. This package belongs to the skill author's ecosystem.
[REMOTE_CODE_EXECUTION]: The skill instructions direct the user to run code directly from the npm registry using npx, which executes remote code in the local environment to provide the skill's primary logic.
[COMMAND_EXECUTION]: The agent is instructed to execute a local Python script scripts/g2b_order_plan.py to perform searches. The script uses standard libraries to handle query parameters and process results.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external government APIs, which could theoretically contain malicious instructions embedded in procurement records.
Ingestion points: The search_order_plans function in scripts/g2b_order_plan.py retrieves data from a remote proxy.
Boundary markers: Results are returned in JSON format, providing structural separation, but there are no specific prompt delimiters to prevent the LLM from following instructions found in data fields like bizNm (business name).
Capability inventory: The skill has the ability to execute network requests to the author's proxy and run shell commands via the CLI wrapper.
Sanitization: The script uses standard JSON parsing, but no content filtering or sanitization of natural language instructions is performed on the data before it is presented to the agent.

g2b-order-plan-search