The Agent Skills Directory

[SAFE]: The skill operates as a static and dynamic analysis tool for MikoPBX source code. All operations, including file reading, grep searches, and Docker interactions, are aligned with its stated purpose of API validation.
[INDIRECT_PROMPT_INJECTION]: The skill ingests PHP source code and system logs (error.log) to perform validation. While this provides a surface for indirect instructions, the skill instructions are focused on structural validation (regex patterns and PHP structure), which reduces the risk of the agent executing arbitrary instructions embedded in the analyzed code.
[COMMAND_EXECUTION]: The skill uses docker exec and curl to interact with the environment. These commands are used for diagnostics (checking environment variables, logs) and testing API behavior. All network targets (PBX_URL) and credentials (TOKEN) are placeholders to be provided by the user, following safe development practices.
[DATA_EXPOSURE]: The skill accesses local project source files and container logs. This access is necessary for the validation task and is performed using standard development tools (find, grep, docker). There are no hardcoded secrets or evidence of unauthorized data exfiltration.

endpoint-validator