The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes various commands using the agkan CLI tool to interact with a task backlog. These commands include agkan task get to retrieve data, agkan task update to modify task content, and agkan task add to create new sub-tasks. These operations are core to the skill's purpose of backlog management.
[PROMPT_INJECTION]: The instructions use strong imperative language (e.g., "MANDATORY", "REQUIRED", "MUST NOT be skipped") to ensure that planning results are persisted back to the task body. While authoritative, this is used to enforce workflow completion rather than to bypass platform safety constraints.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes task descriptions which may contain untrusted data.
Ingestion points: Task data is retrieved via agkan task get <id> --json in the Content Review step.
Boundary markers: There are no delimiters or specific instructions provided to the agent to ignore potential instructions embedded within the retrieved task content.
Capability inventory: The skill has the ability to update tasks (agkan task update), create new tasks (agkan task add), and modify metadata via tags and blocking relationships.
Sanitization: The skill does not perform validation or sanitization of the task body content before it is processed or re-inserted into the backlog via CLI commands.

agkan-planning-subtask