The Agent Skills Directory

[SAFE]: The skill implements a robust validation framework for project planning artifacts. No malicious patterns, obfuscation, or security bypass attempts were detected.
[COMMAND_EXECUTION]: The skill uses local CLI tools (br and bv) to perform task management operations such as creating, updating, and syncing beads (tasks). These operations are restricted to the local project environment and are necessary for the skill's stated purpose of plan validation.
[DATA_EXFILTRATION]: The skill reads project-specific configuration files and artifacts (e.g., CONTEXT.md, phase-contract.md) to perform validation. There are no network operations detected that would allow for the exfiltration of this data to external domains.
[PROMPT_INJECTION]: The instructions incorporate 'Hard Gates' and specific sub-agent prompts to enforce strict logic and quality standards. These prompts do not contain instructions to override AI safety filters or disregard system instructions.

beo-validating