The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on executing local CLI tools such as uv, just, and curl within a specific local directory (/home/vimkim/temp/cubrid-qahome-fetcher). These tools are used for authenticated data retrieval and report generation.
[EXTERNAL_DOWNLOADS]: The skill's primary function is to fetch QA report pages from qahome.cubrid.org. It includes logic to handle authentication and verify that the retrieved content matches the requested build to prevent analysis of stale or incorrect data.
[INDIRECT_PROMPT_INJECTION]: Because the skill processes external HTML content from the QA portal, it possesses an indirect prompt injection surface.
Ingestion points: Raw HTML result pages and summary endpoints (e.g., viewShellTestResult.nhn, showFuntionRes.nhn) fetched from qahome.cubrid.org (SKILL.md).
Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions when the agent processes the fetched HTML content.
Capability inventory: The skill allows shell command execution (uv, just, curl) and file system writes to the runs/ subdirectory (SKILL.md).
Sanitization: There is no mention of sanitizing or escaping the HTML content before it is analyzed by the agent.
[CREDENTIALS_UNSAFE]: The skill demonstrates secure credential management by instructing the agent to read secrets only from environment variables (CUBRID_QA_USER, CUBRID_QA_PASS) rather than hardcoding them or asking the user to provide them in the chat. It explicitly forbids the agent from echoing or logging the password value.

cubrid-qa-fetch