The Agent Skills Directory

[COMMAND_EXECUTION]: The skill frequently executes shell commands via the Bash tool to perform repository reconnaissance, run validation gates (test/lint), and manage git operations. This is a core part of its functionality as a repository management tool.
[EXTERNAL_DOWNLOADS]: The onboard and addon sub-skills facilitate the installation of external coding-agent skills and CLI tools (e.g., DailybotHQ/deepworkplan-skill, DailybotHQ/agent-skill, DailybotHQ/ai-diff-reviewer).
These downloads are performed using npx --yes skills add and are pinned to specific version tags (e.g., @v2.17.0, @v3.10.3).
All referenced external repositories are owned by the skill's authoring organization (DailybotHQ), aligning with the intended product ecosystem.
[REMOTE_CODE_EXECUTION]: By installing and invoking vendored skills from the same vendor, the skill enables a modular execution environment. The installation process uses checksum-verified paths provided by the skills CLI.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to read and process untrusted content from the user's repository (code, docs, tool output). It includes a dedicated 'Untrusted-content rule' in execute/SKILL.md that explicitly instructs the agent to treat such content as data rather than instructions, mitigating the risk of indirect injection.
[DYNAMIC_EXECUTION]: The skill dynamically generates task instructions (N.task_*.md) and repository documentation (AGENTS.md, docs/*.md) based on its analysis of the local tech stack. It also authors 'thin delegator' commands in the .agents/commands/ directory to route alias calls to the installed skill packs.

deepworkplan