The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and act upon data from an external 'inbox' containing tasks (vimarshas) provided by other users or agents.
Ingestion points: Incoming posed_to vimarshas and the AGENTS.md configuration file.
Boundary markers: The skill does not define explicit delimiters or instructions to ignore embedded commands within the vimarshas; it instructs the agent to follow the flow named in the item.
Capability inventory: The agent has the authority to perform git operations (commit, branch, PR merge), modify files ('weaving'), and execute nks_* MCP tools.
Sanitization: No sanitization, validation, or filtering of the incoming task content is described.
[COMMAND_EXECUTION]: The skill requires the agent to perform 'repo rituals' including branch discipline and conventional commits, which involve executing git commands that modify the repository's state.
[PERSISTENCE_MECHANISMS]: The skill establishes a durable 'duty cycle' through 'sleep-poll' mechanisms, ensuring the agent remains active and resumes its cycle across sessions to monitor the repository.

on-duty