The Agent Skills Directory

[SAFE]: The skill implements a comprehensive planning methodology including discovery, synthesis, and task decomposition (beads) for software development projects.
[COMMAND_EXECUTION]: The skill executes node .codex/khuym_status.mjs --json to synchronize repository state, which is a standard procedure for maintaining state within the tool's ecosystem.
[COMMAND_EXECUTION]: The skill utilizes several local CLI tools (br, bv, cass, cm) and standard shell commands (cat, grep) to interact with the repository's task graph and history files.
[EXTERNAL_DOWNLOADS]: The 'Agent D' research instructions allow for the use of web search and fetching tools to gather documentation on third-party libraries and community patterns, which is standard practice for software architecture planning.
[SAFE]: No malicious prompt injection patterns, unauthorized data exfiltration, or persistent access mechanisms were detected in the skill instructions or templates.

planning