The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill uses the ! syntax to execute a shell script yaml-helper.sh at load time. This is used to resolve configuration keys (e.g., automation modes). While this involves command execution during skill initialization, it targets a script within the skill's own directory structure (CLAUDE_SKILL_DIR) and is used for environment setup, which is consistent with the skill's management of project configuration.
[COMMAND_EXECUTION]: The skill uses Bash tools to list directory contents (ls -t) and grep through log files. These operations are standard for log analysis and are limited to the project's log and source directories.
[INDIRECT_PROMPT_INJECTION]: The skill ingests external CI logs (XML and plain text). It processes this data to identify patterns but does not interpret the content as instructions for the agent. The risk is minimized by the use of structured parsing (JUnit/NUnit XML) and specific regex patterns for pass/fail results.
[PRIVILEGE_ESCALATION]: No privilege escalation patterns were detected. The skill operates within standard user permissions and does not use sudo or modify sensitive system paths.
[DATA_EXFILTRATION]: No network operations or attempts to exfiltrate data to external domains were detected.

test-flakiness